Fix MCP lifecycle trace user scope

* feat(agent): thread per-tool reasoning from LLM through to REPL, HTTP, SSE, and DB

Add end-to-end agent reasoning summaries so users can see *why* the
agent chose specific tools, not just what it did.

- Add `reasoning: Option<String>` to `ToolCall` (all providers)
- Populate from LLM response content in `Reasoning::respond_with_tools`
  and `select_tools`, with per-tool override when providers supply it
- Extend `Turn` with `narrative` and `TurnToolCall` with `rationale` +
  `tool_call_id` for identity-based result matching
- Persist reasoning in DB via existing tool_calls JSON (no migration)
- Add `StatusUpdate::ReasoningUpdate` and `SseEvent::ReasoningUpdate` +
  `SseEvent::JobReasoning` for real-time streaming
- Emit reasoning events in both chat dispatcher and worker job path
- Add `/reasoning [N|all]` command for inspecting turn reasoning
- Surface `narrative` and `rationale` in HTTP `/api/chat/history`

Based on the design from #361 and #456, reconstructed cleanly with
Option<String> to minimize blast radius (vs mandatory String that broke
compilation in #456).

Closes #456

Co-Authored-By: panosAthDBX <[email protected]>
Co-Authored-By: Claude Opus 4.6 (1M context) <[email protected]>

* fix: address PR review feedback from Gemini and Copilot

- Fix `_ => Ok(None)` in agent_loop.rs to avoid accidental shutdown
- Fix fallback in record_tool_result_for/record_tool_error_for to use
  first pending call instead of last_mut (parallel execution safety)
- Include per-tool decisions in WASM channel reasoning messages
- Apply truncate_at_tool_tags + clean_response to shared_reasoning in
  select_tools (parity with respond_with_tools)
- Persist turn-level narrative to DB in tool_calls JSON wrapper
- Parse both old (array) and new (object) tool_calls formats in
  build_turns_from_db_messages for backward compatibility
- Populate reasoning from action.reasoning in execute_plan ToolCalls

[skip-regression-check]

Co-Authored-By: Claude Opus 4.6 (1M context) <[email protected]>

* fix: address second round of review comments + merge fixes

- Add reasoning: None to new github_copilot.rs ToolCall sites (from staging merge)
- Run cargo fmt on 4 files with formatting diffs
- Truncate narrative to 1000 chars before DB persistence
- Clone turn data and drop session lock in /reasoning command
- Extract ToolDecisionDto::from_json_array shared helper (deduplicate
  worker/job.rs and orchestrator/api.rs)
- Add unit tests for wrapped tool_calls JSON format with narrative

[skip-regression-check]

Co-Authored-By: Claude Opus 4.6 (1M context) <[email protected]>

* fix: address third round of review comments (Copilot + serrrfirat)

- Reword ToolCall.reasoning docstring to reflect provider-supplied or
  fallback contract
- Sanitize narrative through SafetyLayer before storage/emission
- Clean per-tool reasoning via truncate_at_tool_tags + clean_response
  in select_tools (parity with shared reasoning)
- Convert 4 approval-path recording sites in thread_ops.rs to
  identity-based record_tool_result_for/record_tool_error_for
- Preserve tool_call_id and reasoning through restore_from_messages
- Fix has_result/has_error to reject JSON null values
- Truncate tool_call_id to 128 chars before DB persistence
- Add 4 unit tests for record_tool_result_for/error_for edge cases

Co-Authored-By: Claude Opus 4.6 (1M context) <[email protected]>

* fix: address zmanian review — sanitize JobDelegate reasoning + warn on dropped results

- Sanitize narrative and per-tool rationale through SafetyLayer in
  JobDelegate reasoning events (parity with ChatDelegate)
- Add tracing::warn when record_tool_result_for/error_for drops a
  result because no matching or pending tool call exists
- Add 3 unit tests for reasoning normalization (thinking tags,
  tool tags, empty-after-cleaning)

Co-Authored-By: Claude Opus 4.6 (1M context) <[email protected]>

* fix: address 4 remaining unreplied review comments

- Clean per-tool reasoning in respond_with_tools via truncate_at_tool_tags
  + clean_response (parity with select_tools)
- Handle wrapped JSON format in rebuild_chat_messages_from_db so cold
  hydration works after persist_tool_calls format change
- Update persist_tool_calls doc comment to describe new JSON shape
- Sanitize per-tool rationale through SafetyLayer in ChatDelegate before
  emission and storage (parity with JobDelegate)

Co-Authored-By: Claude Opus 4.6 (1M context) <[email protected]>

* fix: address zmanian review round 2

- Add tracing::debug on fallback-to-pending path in record_tool_result_for
  and record_tool_error_for (item 1)
- Add comment explaining why /reasoning is special-cased in agent_loop.rs
  (item 4)
- Items 2 (narrative persistence), 3 (rationale sanitization), and 5
  (catch-all fix) were already addressed in prior commits

Co-Authored-By: Claude Opus 4.6 (1M context) <[email protected]>

---------

Co-authored-by: panosAthDBX <[email protected]>
Co-authored-by: Claude Opus 4.6 (1M context) <[email protected]>

.github/workflows/e2e.yml

@@ -54,7 +54,7 @@ jobs:
           - group: features
             files: "tests/e2e/scenarios/test_skills.py tests/e2e/scenarios/test_tool_approval.py tests/e2e/scenarios/test_webhook.py"
           - group: extensions
-            files: "tests/e2e/scenarios/test_extensions.py tests/e2e/scenarios/test_extension_oauth.py tests/e2e/scenarios/test_telegram_token_validation.py tests/e2e/scenarios/test_telegram_hot_activation.py tests/e2e/scenarios/test_wasm_lifecycle.py tests/e2e/scenarios/test_tool_execution.py tests/e2e/scenarios/test_pairing.py tests/e2e/scenarios/test_mcp_auth_flow.py tests/e2e/scenarios/test_oauth_credential_fallback.py tests/e2e/scenarios/test_routine_oauth_credential_injection.py"
+            files: "tests/e2e/scenarios/test_extensions.py tests/e2e/scenarios/test_extension_oauth.py tests/e2e/scenarios/test_oauth_url_parameters.py tests/e2e/scenarios/test_telegram_token_validation.py tests/e2e/scenarios/test_telegram_hot_activation.py tests/e2e/scenarios/test_wasm_lifecycle.py tests/e2e/scenarios/test_tool_execution.py tests/e2e/scenarios/test_pairing.py tests/e2e/scenarios/test_mcp_auth_flow.py tests/e2e/scenarios/test_oauth_credential_fallback.py tests/e2e/scenarios/test_routine_oauth_credential_injection.py"
           - group: routines
             files: "tests/e2e/scenarios/test_owner_scope.py tests/e2e/scenarios/test_routine_event_batch.py"
     steps:

.github/workflows/test.yml

@@ -12,6 +12,7 @@ jobs:
   tests:
     name: Tests (${{ matrix.name }})
     runs-on: ubuntu-latest
+    timeout-minutes: 45
     strategy:
       fail-fast: false
       matrix:
@@ -40,11 +41,14 @@ jobs:
       - name: Build WASM channels (for integration tests)
         run: ./scripts/build-wasm-extensions.sh --channels
       - name: Run Tests
-        run: cargo test ${{ matrix.flags }} -- --nocapture
+        run: |
+          timeout --signal=INT --kill-after=30s 40m \
+            cargo test ${{ matrix.flags }} -- --nocapture
 
   heavy-integration-tests:
     name: Heavy Integration Tests
     runs-on: ubuntu-latest
+    timeout-minutes: 20
     steps:
       - name: Checkout repository
         uses: actions/checkout@v6
@@ -58,9 +62,13 @@ jobs:
       - name: Build Telegram WASM channel
         run: cargo build --manifest-path channels-src/telegram/Cargo.toml --target wasm32-wasip2 --release
       - name: Run thread scheduling integration tests
-        run: cargo test --no-default-features --features libsql,integration --test e2e_thread_scheduling -- --nocapture
+        run: |
+          timeout --signal=INT --kill-after=30s 15m \
+            cargo test --no-default-features --features libsql,integration --test e2e_thread_scheduling -- --nocapture
       - name: Run Telegram thread-scope regression test
-        run: cargo test --features integration --test telegram_auth_integration test_private_messages_use_chat_id_as_thread_scope -- --exact
+        run: |
+          timeout --signal=INT --kill-after=30s 10m \
+            cargo test --features integration --test telegram_auth_integration test_private_messages_use_chat_id_as_thread_scope -- --exact
 
   telegram-tests:
     name: Telegram Channel Tests
@@ -68,6 +76,7 @@ jobs:
       github.event_name != 'pull_request' ||
       github.base_ref != 'staging'
     runs-on: ubuntu-latest
+    timeout-minutes: 15
     steps:
       - name: Checkout repository
         uses: actions/checkout@v6
@@ -75,7 +84,9 @@ jobs:
         uses: dtolnay/rust-toolchain@stable
       - uses: Swatinem/rust-cache@v2
       - name: Run Telegram Channel Tests
-        run: cargo test --manifest-path channels-src/telegram/Cargo.toml -- --nocapture
+        run: |
+          timeout --signal=INT --kill-after=30s 10m \
+            cargo test --manifest-path channels-src/telegram/Cargo.toml -- --nocapture
 
   windows-build:
     name: Windows Build (${{ matrix.name }})
@@ -110,6 +121,7 @@ jobs:
       github.event_name != 'pull_request' ||
       github.base_ref != 'staging'
     runs-on: ubuntu-latest
+    timeout-minutes: 30
     steps:
       - name: Checkout repository
         uses: actions/checkout@v6
@@ -125,7 +137,9 @@ jobs:
       - name: Build all WASM extensions against current WIT
         run: ./scripts/build-wasm-extensions.sh
       - name: Instantiation test (host linker compatibility)
-        run: cargo test --all-features wit_compat -- --nocapture
+        run: |
+          timeout --signal=INT --kill-after=30s 20m \
+            cargo test --all-features wit_compat -- --nocapture
 
   bench-compile:
     name: Benchmark Compilation

Cargo.lock

@@ -157,7 +157,7 @@ version = "1.1.5"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "40c48f72fd53cd289104fc64099abca73db4166ad86ea0b4341abe65af83dadc"
 dependencies = [
- "windows-sys 0.61.2",
+ "windows-sys 0.60.2",
 ]
 
 [[package]]
@@ -168,7 +168,7 @@ checksum = "291e6a250ff86cd4a820112fb8898808a366d8f9f58ce16d1f538353ad55747d"
 dependencies = [
  "anstyle",
  "once_cell_polyfill",
- "windows-sys 0.61.2",
+ "windows-sys 0.60.2",
 ]
 
 [[package]]
@@ -2136,7 +2136,7 @@ dependencies = [
  "libc",
  "option-ext",
  "redox_users 0.5.2",
- "windows-sys 0.61.2",
+ "windows-sys 0.59.0",
 ]
 
 [[package]]
@@ -2323,7 +2323,7 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "39cab71617ae0d63f51a36d69f866391735b51691dbda63cf6f96d042b63efeb"
 dependencies = [
  "libc",
- "windows-sys 0.61.2",
+ "windows-sys 0.52.0",
 ]
 
 [[package]]
@@ -3428,6 +3428,7 @@ dependencies = [
  "hyper-util",
  "iana-time-zone",
  "insta",
+ "ironclaw_common",
  "ironclaw_safety",
  "json5",
  "libsql",
@@ -3485,6 +3486,14 @@ dependencies = [
  "zip",
 ]
 
+[[package]]
+name = "ironclaw_common"
+version = "0.1.0"
+dependencies = [
+ "serde",
+ "serde_json",
+]
+
 [[package]]
 name = "ironclaw_safety"
 version = "0.1.0"
@@ -4134,7 +4143,7 @@ version = "0.50.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "7957b9740744892f114936ab4a57b3f487491bbeafaf8083688b16841a4240e5"
 dependencies = [
- "windows-sys 0.61.2",
+ "windows-sys 0.59.0",
 ]
 
 [[package]]
@@ -5472,7 +5481,7 @@ dependencies = [
  "errno",
  "libc",
  "linux-raw-sys 0.12.1",
- "windows-sys 0.61.2",
+ "windows-sys 0.52.0",
 ]
 
 [[package]]
@@ -6154,7 +6163,7 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "3a766e1110788c36f4fa1c2b71b387a7815aa65f88ce0229841826633d93723e"
 dependencies = [
  "libc",
- "windows-sys 0.61.2",
+ "windows-sys 0.60.2",
 ]
 
 [[package]]
@@ -6354,9 +6363,9 @@ checksum = "55937e1799185b12863d447f42597ed69d9928686b8d88a1df17376a097d8369"
 
 [[package]]
 name = "tar"
-version = "0.4.44"
+version = "0.4.45"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "1d863878d212c87a19c1a610eb53bb01fe12951c0501cf5a0d65f724914a667a"
+checksum = "22692a6476a21fa75fdfc11d452fda482af402c008cdbaf3476414e122040973"
 dependencies = [
  "filetime",
  "libc",
@@ -6379,7 +6388,7 @@ dependencies = [
  "getrandom 0.4.2",
  "once_cell",
  "rustix 1.1.4",
- "windows-sys 0.61.2",
+ "windows-sys 0.52.0",
 ]
 
 [[package]]
@@ -7179,7 +7188,7 @@ checksum = "f2f6fb2847f6742cd76af783a2a2c49e9375d0a111c7bef6f71cd9e738c72d6e"
 dependencies = [
  "memoffset",
  "tempfile",
- "windows-sys 0.61.2",
+ "windows-sys 0.60.2",
 ]
 
 [[package]]
@@ -8029,7 +8038,7 @@ version = "0.1.11"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "c2a7b1c03c876122aa43f3020e6c3c3ee5c05081c9a00739faf7503aeba10d22"
 dependencies = [
- "windows-sys 0.61.2",
+ "windows-sys 0.48.0",
 ]
 
 [[package]]

Cargo.toml

@@ -1,5 +1,5 @@
 [workspace]
-members = [".", "crates/ironclaw_safety"]
+members = [".", "crates/ironclaw_common", "crates/ironclaw_safety"]
 exclude = [
     "channels-src/discord",
     "channels-src/telegram",
@@ -100,6 +100,9 @@ tower-http = { version = "0.6", features = ["trace", "cors", "set-header"] }
 # Cron scheduling for routines
 cron = "0.13"
 
+# Shared types
+ironclaw_common = { path = "crates/ironclaw_common", version = "0.1.0" }
+
 # Safety/sanitization
 ironclaw_safety = { path = "crates/ironclaw_safety", version = "0.1.0" }
 regex = "1"

FEATURE_PARITY.md

@@ -161,7 +161,7 @@ This document tracks feature parity between IronClaw (Rust implementation) and O
 | `config` | ✅ | ✅ | - | Read/write config plus validate/path helpers |
 | `backup` | ✅ | ❌ | P3 | Create/verify local backup archives |
 | `channels` | ✅ | 🚧 | P2 | `list` implemented; `enable`/`disable`/`status` deferred pending config source unification |
-| `models` | ✅ | 🚧 | - | Model selector in TUI |
+| `models` | ✅ | 🚧 | P1 | `models list [<provider>]` (`--verbose`, `--json`; fetches live model list when provider specified), `models status` (`--json`), `models set <model>`, `models set-provider <provider> [--model model]` (alias normalization, config.toml + .env persistence). Remaining: `set` doesn't validate model against live list. |
 | `status` | ✅ | ✅ | - | System status (enriched session details) |
 | `agents` | ✅ | ❌ | P3 | Multi-agent management |
 | `sessions` | ✅ | ❌ | P3 | Session listing (shows subagent models) |

channels-src/feishu/feishu.capabilities.json

@@ -21,7 +21,7 @@
       },
       {
         "name": "feishu_app_secret",
-        "prompt": "Enter your Feishu/Lark App Secret",
+        "prompt": "Enter your Feishu/Lark App Secret (from your app settings at open.feishu.cn)",
         "optional": false
       },
       {

crates/ironclaw_common/Cargo.toml

@@ -0,0 +1,18 @@
+[package]
+name = "ironclaw_common"
+version = "0.1.0"
+edition = "2024"
+rust-version = "1.92"
+description = "Shared types and utilities for the IronClaw workspace"
+authors = ["NEAR AI <[email protected]>"]
+license = "MIT OR Apache-2.0"
+homepage = "https://github.com/nearai/ironclaw"
+repository = "https://github.com/nearai/ironclaw"
+publish = false
+
+[package.metadata.dist]
+dist = false
+
+[dependencies]
+serde = { version = "1", features = ["derive"] }
+serde_json = "1"

crates/ironclaw_common/src/event.rs

@@ -0,0 +1,393 @@
+//! Application-wide event types.
+//!
+//! `AppEvent` is the real-time event protocol used across the entire
+//! application.  The web gateway serialises these to SSE / WebSocket
+//! frames, but other subsystems (agent loop, orchestrator, extensions)
+//! produce and consume them too.
+
+use serde::{Deserialize, Serialize};
+
+/// A single tool decision in a reasoning update (SSE DTO).
+#[derive(Debug, Clone, Serialize, Deserialize)]
+pub struct ToolDecisionDto {
+    pub tool_name: String,
+    pub rationale: String,
+}
+
+impl ToolDecisionDto {
+    /// Parse a list of tool decisions from a JSON array value.
+    pub fn from_json_array(value: &serde_json::Value) -> Vec<Self> {
+        value
+            .as_array()
+            .map(|arr| {
+                arr.iter()
+                    .filter_map(|d| {
+                        Some(Self {
+                            tool_name: d.get("tool_name")?.as_str()?.to_string(),
+                            rationale: d.get("rationale")?.as_str()?.to_string(),
+                        })
+                    })
+                    .collect()
+            })
+            .unwrap_or_default()
+    }
+}
+
+#[derive(Debug, Clone, Serialize, Deserialize)]
+#[serde(tag = "type")]
+pub enum AppEvent {
+    #[serde(rename = "response")]
+    Response { content: String, thread_id: String },
+    #[serde(rename = "thinking")]
+    Thinking {
+        message: String,
+        #[serde(skip_serializing_if = "Option::is_none")]
+        thread_id: Option<String>,
+    },
+    #[serde(rename = "tool_started")]
+    ToolStarted {
+        name: String,
+        #[serde(skip_serializing_if = "Option::is_none")]
+        thread_id: Option<String>,
+    },
+    #[serde(rename = "tool_completed")]
+    ToolCompleted {
+        name: String,
+        success: bool,
+        #[serde(skip_serializing_if = "Option::is_none")]
+        error: Option<String>,
+        #[serde(skip_serializing_if = "Option::is_none")]
+        parameters: Option<String>,
+        #[serde(skip_serializing_if = "Option::is_none")]
+        thread_id: Option<String>,
+    },
+    #[serde(rename = "tool_result")]
+    ToolResult {
+        name: String,
+        preview: String,
+        #[serde(skip_serializing_if = "Option::is_none")]
+        thread_id: Option<String>,
+    },
+    #[serde(rename = "stream_chunk")]
+    StreamChunk {
+        content: String,
+        #[serde(skip_serializing_if = "Option::is_none")]
+        thread_id: Option<String>,
+    },
+    #[serde(rename = "status")]
+    Status {
+        message: String,
+        #[serde(skip_serializing_if = "Option::is_none")]
+        thread_id: Option<String>,
+    },
+    #[serde(rename = "job_started")]
+    JobStarted {
+        job_id: String,
+        title: String,
+        browse_url: String,
+    },
+    #[serde(rename = "approval_needed")]
+    ApprovalNeeded {
+        request_id: String,
+        tool_name: String,
+        description: String,
+        parameters: String,
+        #[serde(skip_serializing_if = "Option::is_none")]
+        thread_id: Option<String>,
+        /// Whether the "always" auto-approve option should be shown.
+        allow_always: bool,
+    },
+    #[serde(rename = "auth_required")]
+    AuthRequired {
+        extension_name: String,
+        #[serde(skip_serializing_if = "Option::is_none")]
+        instructions: Option<String>,
+        #[serde(skip_serializing_if = "Option::is_none")]
+        auth_url: Option<String>,
+        #[serde(skip_serializing_if = "Option::is_none")]
+        setup_url: Option<String>,
+    },
+    #[serde(rename = "auth_completed")]
+    AuthCompleted {
+        extension_name: String,
+        success: bool,
+        message: String,
+    },
+    #[serde(rename = "error")]
+    Error {
+        message: String,
+        #[serde(skip_serializing_if = "Option::is_none")]
+        thread_id: Option<String>,
+    },
+    #[serde(rename = "heartbeat")]
+    Heartbeat,
+
+    // Sandbox job streaming events (worker + Claude Code bridge)
+    #[serde(rename = "job_message")]
+    JobMessage {
+        job_id: String,
+        role: String,
+        content: String,
+    },
+    #[serde(rename = "job_tool_use")]
+    JobToolUse {
+        job_id: String,
+        tool_name: String,
+        input: serde_json::Value,
+    },
+    #[serde(rename = "job_tool_result")]
+    JobToolResult {
+        job_id: String,
+        tool_name: String,
+        output: String,
+    },
+    #[serde(rename = "job_status")]
+    JobStatus { job_id: String, message: String },
+    #[serde(rename = "job_result")]
+    JobResult {
+        job_id: String,
+        status: String,
+        #[serde(skip_serializing_if = "Option::is_none")]
+        session_id: Option<String>,
+        #[serde(skip_serializing_if = "Option::is_none")]
+        fallback_deliverable: Option<serde_json::Value>,
+    },
+
+    /// An image was generated by a tool.
+    #[serde(rename = "image_generated")]
+    ImageGenerated {
+        data_url: String,
+        #[serde(skip_serializing_if = "Option::is_none")]
+        path: Option<String>,
+        #[serde(skip_serializing_if = "Option::is_none")]
+        thread_id: Option<String>,
+    },
+
+    /// Suggested follow-up messages for the user.
+    #[serde(rename = "suggestions")]
+    Suggestions {
+        suggestions: Vec<String>,
+        #[serde(skip_serializing_if = "Option::is_none")]
+        thread_id: Option<String>,
+    },
+
+    /// Per-turn token usage and cost summary.
+    #[serde(rename = "turn_cost")]
+    TurnCost {
+        input_tokens: u64,
+        output_tokens: u64,
+        cost_usd: String,
+        #[serde(skip_serializing_if = "Option::is_none")]
+        thread_id: Option<String>,
+    },
+
+    /// Extension activation status change (WASM channels).
+    #[serde(rename = "extension_status")]
+    ExtensionStatus {
+        extension_name: String,
+        status: String,
+        #[serde(skip_serializing_if = "Option::is_none")]
+        message: Option<String>,
+    },
+
+    /// Agent reasoning update (why it chose specific tools).
+    #[serde(rename = "reasoning_update")]
+    ReasoningUpdate {
+        narrative: String,
+        decisions: Vec<ToolDecisionDto>,
+        #[serde(skip_serializing_if = "Option::is_none")]
+        thread_id: Option<String>,
+    },
+
+    /// Reasoning update for a sandbox job.
+    #[serde(rename = "job_reasoning")]
+    JobReasoning {
+        job_id: String,
+        narrative: String,
+        decisions: Vec<ToolDecisionDto>,
+    },
+}
+
+impl AppEvent {
+    /// The wire-format event type string (matches the `#[serde(rename)]` value).
+    pub fn event_type(&self) -> &'static str {
+        match self {
+            Self::Response { .. } => "response",
+            Self::Thinking { .. } => "thinking",
+            Self::ToolStarted { .. } => "tool_started",
+            Self::ToolCompleted { .. } => "tool_completed",
+            Self::ToolResult { .. } => "tool_result",
+            Self::StreamChunk { .. } => "stream_chunk",
+            Self::Status { .. } => "status",
+            Self::JobStarted { .. } => "job_started",
+            Self::ApprovalNeeded { .. } => "approval_needed",
+            Self::AuthRequired { .. } => "auth_required",
+            Self::AuthCompleted { .. } => "auth_completed",
+            Self::Error { .. } => "error",
+            Self::Heartbeat => "heartbeat",
+            Self::JobMessage { .. } => "job_message",
+            Self::JobToolUse { .. } => "job_tool_use",
+            Self::JobToolResult { .. } => "job_tool_result",
+            Self::JobStatus { .. } => "job_status",
+            Self::JobResult { .. } => "job_result",
+            Self::ImageGenerated { .. } => "image_generated",
+            Self::Suggestions { .. } => "suggestions",
+            Self::TurnCost { .. } => "turn_cost",
+            Self::ExtensionStatus { .. } => "extension_status",
+            Self::ReasoningUpdate { .. } => "reasoning_update",
+            Self::JobReasoning { .. } => "job_reasoning",
+        }
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    /// Verify that `event_type()` returns the same string as the serde
+    /// `"type"` field for every variant.  This catches drift between the
+    /// `#[serde(rename)]` attributes and the manual match arms.
+    #[test]
+    fn event_type_matches_serde_type_field() {
+        let variants: Vec<AppEvent> = vec![
+            AppEvent::Response {
+                content: String::new(),
+                thread_id: String::new(),
+            },
+            AppEvent::Thinking {
+                message: String::new(),
+                thread_id: None,
+            },
+            AppEvent::ToolStarted {
+                name: String::new(),
+                thread_id: None,
+            },
+            AppEvent::ToolCompleted {
+                name: String::new(),
+                success: true,
+                error: None,
+                parameters: None,
+                thread_id: None,
+            },
+            AppEvent::ToolResult {
+                name: String::new(),
+                preview: String::new(),
+                thread_id: None,
+            },
+            AppEvent::StreamChunk {
+                content: String::new(),
+                thread_id: None,
+            },
+            AppEvent::Status {
+                message: String::new(),
+                thread_id: None,
+            },
+            AppEvent::JobStarted {
+                job_id: String::new(),
+                title: String::new(),
+                browse_url: String::new(),
+            },
+            AppEvent::ApprovalNeeded {
+                request_id: String::new(),
+                tool_name: String::new(),
+                description: String::new(),
+                parameters: String::new(),
+                thread_id: None,
+                allow_always: false,
+            },
+            AppEvent::AuthRequired {
+                extension_name: String::new(),
+                instructions: None,
+                auth_url: None,
+                setup_url: None,
+            },
+            AppEvent::AuthCompleted {
+                extension_name: String::new(),
+                success: true,
+                message: String::new(),
+            },
+            AppEvent::Error {
+                message: String::new(),
+                thread_id: None,
+            },
+            AppEvent::Heartbeat,
+            AppEvent::JobMessage {
+                job_id: String::new(),
+                role: String::new(),
+                content: String::new(),
+            },
+            AppEvent::JobToolUse {
+                job_id: String::new(),
+                tool_name: String::new(),
+                input: serde_json::Value::Null,
+            },
+            AppEvent::JobToolResult {
+                job_id: String::new(),
+                tool_name: String::new(),
+                output: String::new(),
+            },
+            AppEvent::JobStatus {
+                job_id: String::new(),
+                message: String::new(),
+            },
+            AppEvent::JobResult {
+                job_id: String::new(),
+                status: String::new(),
+                session_id: None,
+                fallback_deliverable: None,
+            },
+            AppEvent::ImageGenerated {
+                data_url: String::new(),
+                path: None,
+                thread_id: None,
+            },
+            AppEvent::Suggestions {
+                suggestions: vec![],
+                thread_id: None,
+            },
+            AppEvent::TurnCost {
+                input_tokens: 0,
+                output_tokens: 0,
+                cost_usd: String::new(),
+                thread_id: None,
+            },
+            AppEvent::ExtensionStatus {
+                extension_name: String::new(),
+                status: String::new(),
+                message: None,
+            },
+            AppEvent::ReasoningUpdate {
+                narrative: String::new(),
+                decisions: vec![],
+                thread_id: None,
+            },
+            AppEvent::JobReasoning {
+                job_id: String::new(),
+                narrative: String::new(),
+                decisions: vec![],
+            },
+        ];
+
+        for variant in &variants {
+            let json: serde_json::Value = serde_json::to_value(variant).unwrap();
+            let serde_type = json["type"].as_str().unwrap();
+            assert_eq!(
+                variant.event_type(),
+                serde_type,
+                "event_type() mismatch for variant: {:?}",
+                variant
+            );
+        }
+    }
+
+    #[test]
+    fn round_trip_deserialize() {
+        let original = AppEvent::Response {
+            content: "hello".to_string(),
+            thread_id: "t1".to_string(),
+        };
+        let json = serde_json::to_string(&original).unwrap();
+        let deserialized: AppEvent = serde_json::from_str(&json).unwrap();
+        assert_eq!(deserialized.event_type(), "response");
+    }
+}

crates/ironclaw_common/src/lib.rs

@@ -0,0 +1,7 @@
+//! Shared types and utilities for the IronClaw workspace.
+
+mod event;
+mod util;
+
+pub use event::{AppEvent, ToolDecisionDto};
+pub use util::truncate_preview;

crates/ironclaw_common/src/util.rs

@@ -0,0 +1,100 @@
+//! Shared utility functions.
+
+/// Truncate a string to at most `max_bytes` bytes at a char boundary, appending "...".
+///
+/// If the input is wrapped in `<tool_output ...>...</tool_output>` and truncation
+/// removes the closing tag, the tag is re-appended so downstream XML parsers
+/// never see an unclosed element.
+pub fn truncate_preview(s: &str, max_bytes: usize) -> String {
+    if s.len() <= max_bytes {
+        return s.to_string();
+    }
+    // Walk backwards from max_bytes to find a valid char boundary
+    let mut end = max_bytes;
+    while end > 0 && !s.is_char_boundary(end) {
+        end -= 1;
+    }
+    let mut result = format!("{}...", &s[..end]);
+
+    // Re-close <tool_output> if truncation cut through the closing tag.
+    if s.starts_with("<tool_output") && !result.ends_with("</tool_output>") {
+        result.push_str("\n</tool_output>");
+    }
+
+    result
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn test_truncate_preview_short_string() {
+        assert_eq!(truncate_preview("hello", 10), "hello");
+    }
+
+    #[test]
+    fn test_truncate_preview_exact_boundary() {
+        assert_eq!(truncate_preview("hello", 5), "hello");
+    }
+
+    #[test]
+    fn test_truncate_preview_truncates_ascii() {
+        assert_eq!(truncate_preview("hello world", 5), "hello...");
+    }
+
+    #[test]
+    fn test_truncate_preview_empty_string() {
+        assert_eq!(truncate_preview("", 10), "");
+    }
+
+    #[test]
+    fn test_truncate_preview_multibyte_char_boundary() {
+        let s = "a\u{20AC}b";
+        let result = truncate_preview(s, 3);
+        assert_eq!(result, "a...");
+    }
+
+    #[test]
+    fn test_truncate_preview_emoji() {
+        let s = "hi\u{1F980}";
+        let result = truncate_preview(s, 4);
+        assert_eq!(result, "hi...");
+    }
+
+    #[test]
+    fn test_truncate_preview_cjk() {
+        let s = "\u{4F60}\u{597D}\u{4E16}\u{754C}";
+        let result = truncate_preview(s, 7);
+        assert_eq!(result, "\u{4F60}\u{597D}...");
+    }
+
+    #[test]
+    fn test_truncate_preview_zero_max_bytes() {
+        assert_eq!(truncate_preview("hello", 0), "...");
+    }
+
+    #[test]
+    fn test_truncate_preview_closes_tool_output_tag() {
+        let s = "<tool_output name=\"search\">\nSome very long content here\n</tool_output>";
+        let result = truncate_preview(s, 60);
+        assert!(result.ends_with("</tool_output>"));
+        assert!(result.contains("..."));
+    }
+
+    #[test]
+    fn test_truncate_preview_no_extra_close_when_intact() {
+        let s = "<tool_output name=\"echo\">\nshort\n</tool_output>";
+        let result = truncate_preview(s, 500);
+        assert_eq!(result, s);
+        assert_eq!(result.matches("</tool_output>").count(), 1);
+    }
+
+    #[test]
+    fn test_truncate_preview_non_xml_unaffected() {
+        let s = "Just a plain long string that gets truncated";
+        let result = truncate_preview(s, 10);
+        assert_eq!(result, "Just a pla...");
+        assert!(!result.contains("</tool_output>"));
+    }
+}

src/agent/agent_loop.rs

@@ -16,6 +16,7 @@ use crate::agent::context_monitor::ContextMonitor;
 use crate::agent::heartbeat::spawn_heartbeat;
 use crate::agent::routine_engine::{RoutineEngine, spawn_cron_ticker};
 use crate::agent::self_repair::{DefaultSelfRepair, RepairResult, SelfRepair};
+use crate::agent::session::ThreadState;
 use crate::agent::session_manager::SessionManager;
 use crate::agent::submission::{Submission, SubmissionParser, SubmissionResult};
 use crate::agent::{HeartbeatConfig as AgentHeartbeatConfig, Router, Scheduler, SchedulerDeps};
@@ -84,6 +85,15 @@ fn resolve_owner_scope_notification_user(
     trimmed_option(explicit_user).or_else(|| trimmed_option(owner_fallback))
 }
 
+fn is_single_message_repl(message: &IncomingMessage) -> bool {
+    message.channel == "repl"
+        && message
+            .metadata
+            .get("single_message_mode")
+            .and_then(|value| value.as_bool())
+            .unwrap_or(false)
+}
+
 async fn resolve_channel_notification_user(
     extension_manager: Option<&Arc<ExtensionManager>>,
     channel: Option<&str>,
@@ -157,8 +167,8 @@ pub struct AgentDeps {
     pub hooks: Arc<HookRegistry>,
     /// Cost enforcement guardrails (daily budget, hourly rate limits).
     pub cost_guard: Arc<crate::agent::cost_guard::CostGuard>,
-    /// SSE broadcast sender for live job event streaming to the web gateway.
+    /// SSE manager for live job event streaming to the web gateway.
-    pub sse_tx: Option<tokio::sync::broadcast::Sender<crate::channels::web::types::SseEvent>>,
+    pub sse_tx: Option<Arc<crate::channels::web::sse::SseManager>>,
     /// HTTP interceptor for trace recording/replay.
     pub http_interceptor: Option<Arc<dyn crate::llm::recording::HttpInterceptor>>,
     /// Audio transcription middleware for voice messages.
@@ -169,6 +179,9 @@ pub struct AgentDeps {
     pub sandbox_readiness: crate::agent::routine_engine::SandboxReadiness,
     /// Software builder for self-repair tool rebuilding.
     pub builder: Option<Arc<dyn crate::tools::SoftwareBuilder>>,
+    /// Resolved LLM backend identifier (e.g., "nearai", "openai", "groq").
+    /// Used by `/model` persistence to determine which env var to update.
+    pub llm_backend: String,
 }
 
 /// The main agent that coordinates all components.
@@ -235,8 +248,8 @@ impl Agent {
                 hooks: deps.hooks.clone(),
             },
         );
-        if let Some(ref tx) = deps.sse_tx {
+        if let Some(ref sse) = deps.sse_tx {
-            scheduler.set_sse_sender(tx.clone());
+            scheduler.set_sse_sender(Arc::clone(sse));
         }
         if let Some(ref interceptor) = deps.http_interceptor {
             scheduler.set_http_interceptor(Arc::clone(interceptor));
@@ -1052,10 +1065,11 @@ impl Agent {
             } else {
                 drop(sess);
                 self.session_manager
-                    .resolve_thread(
+                    .resolve_thread_with_parsed_uuid(
                         &message.user_id,
                         &message.channel,
                         message.conversation_scope(),
+                        approval_thread_uuid,
                     )
                     .await
             }
@@ -1136,9 +1150,14 @@ impl Agent {
             && let Submission::UserInput { ref content } = submission
             && let Some(engine) = self.routine_engine().await
         {
-            let fired = engine
+            let single_message_repl = is_single_message_repl(message);
-                .check_event_triggers(&message.user_id, &message.channel, content)
+            // Use post-hook content so that BeforeInbound hooks that rewrite
-                .await;
+            // input are respected by event trigger matching.
+            let fired = if single_message_repl {
+                engine.check_event_triggers_and_wait(message, content).await
+            } else {
+                engine.check_event_triggers(message, content).await
+            };
             if fired > 0 {
                 tracing::debug!(
                     channel = %message.channel,
@@ -1146,10 +1165,16 @@ impl Agent {
                     fired,
                     "Consumed inbound user message with matching event-triggered routine(s)"
                 );
-                return Ok(Some(String::new()));
+                return if single_message_repl {
+                    Ok(None)
+                } else {
+                    Ok(Some(String::new()))
+                };
             }
         }
 
+        let session_for_empty_exit = Arc::clone(&session);
+
         // Process based on submission type
         let result = match submission {
             Submission::UserInput { content } => {
@@ -1246,6 +1271,28 @@ impl Agent {
                     command,
                     message.channel
                 );
+                // /reasoning is special-cased here (not in handle_system_command)
+                // because it needs the session + thread_id to read turn reasoning
+                // data, which handle_system_command's signature doesn't provide.
+                if command == "reasoning" {
+                    let result = self
+                        .handle_reasoning_command(&args, &session, thread_id)
+                        .await;
+                    return match result {
+                        SubmissionResult::Response { content } => Ok(Some(content)),
+                        SubmissionResult::Ok { message } => Ok(message),
+                        SubmissionResult::Error { message } => {
+                            Ok(Some(format!("Error: {}", message)))
+                        }
+                        _ => {
+                            if is_single_message_repl(message) {
+                                Ok(None)
+                            } else {
+                                Ok(Some(String::new()))
+                            }
+                        }
+                    };
+                }
                 // Authorization checks (including restart channel check) are enforced in handle_system_command
                 self.handle_system_command(&command, &args, &message.channel)
                     .await
@@ -1305,7 +1352,26 @@ impl Agent {
                     Ok(Some(content))
                 }
             }
-            SubmissionResult::Ok { message } => Ok(message),
+            SubmissionResult::Ok {
+                message: output_message,
+            } => {
+                let should_exit =
+                    if output_message.as_deref() == Some("") && is_single_message_repl(message) {
+                        let sess = session_for_empty_exit.lock().await;
+                        sess.threads
+                            .get(&thread_id)
+                            .map(|thread| thread.state != ThreadState::AwaitingApproval)
+                            .unwrap_or(true)
+                    } else {
+                        false
+                    };
+
+                if should_exit {
+                    Ok(None)
+                } else {
+                    Ok(output_message)
+                }
+            }
             SubmissionResult::Error { message } => Ok(Some(format!("Error: {}", message))),
             SubmissionResult::Interrupted => Ok(Some("Interrupted.".into())),
             SubmissionResult::NeedApproval { .. } => {
@@ -1321,7 +1387,7 @@ impl Agent {
 #[cfg(test)]
 mod tests {
     use super::{
-        chat_tool_execution_metadata, resolve_routine_notification_user,
+        chat_tool_execution_metadata, is_single_message_repl, resolve_routine_notification_user,
         should_fallback_routine_notification, truncate_for_preview,
     };
     use crate::channels::IncomingMessage;
@@ -1483,4 +1549,17 @@ mod tests {
 
         assert!(should_fallback_routine_notification(&error)); // safety: test-only assertion
     }
+
+    #[test]
+    fn single_message_repl_detection_requires_repl_channel_and_metadata_flag() {
+        let repl = IncomingMessage::new("repl", "owner-scope", "hello")
+            .with_metadata(serde_json::json!({ "single_message_mode": true }));
+        let gateway = IncomingMessage::new("gateway", "owner-scope", "hello")
+            .with_metadata(serde_json::json!({ "single_message_mode": true }));
+        let plain_repl = IncomingMessage::new("repl", "owner-scope", "hello");
+
+        assert!(is_single_message_repl(&repl)); // safety: test-only assertion
+        assert!(!is_single_message_repl(&gateway)); // safety: test-only assertion
+        assert!(!is_single_message_repl(&plain_repl)); // safety: test-only assertion
+    }
 }

src/agent/agentic_loop.rs

@@ -414,6 +414,7 @@ mod tests {
             id: "call_1".to_string(),
             name: "echo".to_string(),
             arguments: serde_json::json!({}),
+            reasoning: None,
         };
         let delegate = MockDelegate::new(vec![
             tool_calls_output(vec![tool_call]),

src/agent/commands.rs

@@ -465,6 +465,94 @@ impl Agent {
         }
     }
 
+    /// Handle `/reasoning [N|all]` — show reasoning history for the active thread.
+    pub(super) async fn handle_reasoning_command(
+        &self,
+        args: &[String],
+        session: &Arc<Mutex<Session>>,
+        thread_id: Uuid,
+    ) -> SubmissionResult {
+        // Clone the turn data we need, then drop the session lock.
+        let turns_snapshot: Vec<(
+            usize,
+            Option<String>,
+            Vec<crate::agent::session::TurnToolCall>,
+        )>;
+        {
+            let sess = session.lock().await;
+            let thread = match sess.threads.get(&thread_id) {
+                Some(t) => t,
+                None => return SubmissionResult::error("No active thread."),
+            };
+
+            if thread.turns.is_empty() {
+                return SubmissionResult::ok_with_message("No turns yet.");
+            }
+
+            // Parse argument: default=last turn, "all"=all turns, N=specific turn (1-based).
+            let selected: Vec<&crate::agent::session::Turn> = match args.first().map(|s| s.as_str())
+            {
+                Some("all") => thread.turns.iter().collect(),
+                Some(n) => match n.parse::<usize>() {
+                    Ok(0) => return SubmissionResult::error("Turn numbers start at 1."),
+                    Ok(num) if num > thread.turns.len() => {
+                        return SubmissionResult::error(format!(
+                            "Turn {} does not exist (max: {}).",
+                            num,
+                            thread.turns.len()
+                        ));
+                    }
+                    Ok(num) => vec![&thread.turns[num - 1]],
+                    Err(_) => return SubmissionResult::error("Usage: /reasoning [N|all]"),
+                },
+                None => {
+                    // Default: last turn that has tool calls
+                    match thread.turns.iter().rev().find(|t| !t.tool_calls.is_empty()) {
+                        Some(t) => vec![t],
+                        None => {
+                            return SubmissionResult::ok_with_message("No turns with tool calls.");
+                        }
+                    }
+                }
+            };
+
+            turns_snapshot = selected
+                .into_iter()
+                .map(|t| (t.turn_number, t.narrative.clone(), t.tool_calls.clone()))
+                .collect();
+        }
+        // Session lock is now dropped — format output without holding it.
+
+        let mut output = String::new();
+        for (turn_number, narrative, tool_calls) in &turns_snapshot {
+            output.push_str(&format!("--- Turn {} ---\n", turn_number + 1));
+            if let Some(narrative) = narrative {
+                output.push_str(&format!("Reasoning: {}\n", narrative));
+            }
+            if tool_calls.is_empty() {
+                output.push_str("  (no tool calls)\n");
+            } else {
+                for tc in tool_calls {
+                    let status = if tc.error.is_some() {
+                        "error"
+                    } else if tc.result.is_some() {
+                        "ok"
+                    } else {
+                        "pending"
+                    };
+                    output.push_str(&format!("  {} [{}]", tc.name, status));
+                    if let Some(ref rationale) = tc.rationale {
+                        output.push_str(&format!(" — {}", rationale));
+                    }
+                    output.push('\n');
+                }
+            }
+            output.push('\n');
+        }
+
+        SubmissionResult::response(output.trim_end())
+    }
+
     /// Handle system commands that bypass thread-state checks entirely.
     pub(super) async fn handle_system_command(
         &self,
@@ -480,6 +568,7 @@ impl Agent {
                 "  /version          Show version info\n",
                 "  /tools            List available tools\n",
                 "  /debug            Toggle debug mode\n",
+                "  /reasoning [N|all] Show agent reasoning for turns\n",
                 "  /ping             Connectivity check\n",
                 "\n",
                 "Jobs:\n",
@@ -841,12 +930,50 @@ impl Agent {
                 .await
             {
                 tracing::warn!("Failed to persist model to DB: {}", e);
+            } else {
+                tracing::debug!("Persisted selected_model to DB: {}", model);
             }
+        } else {
+            tracing::warn!("No database store available — model choice will not persist to DB");
         }
 
-        // 2. Update TOML config file if it exists (sync I/O in spawn_blocking).
+        // 2. Update .env and TOML config file (sync I/O in spawn_blocking).
         let model_owned = model.to_string();
+        let backend = self.deps.llm_backend.clone();
         if let Err(e) = tokio::task::spawn_blocking(move || {
+            // 2a. Update the backend-specific model env var in ~/.ironclaw/.env.
+            //
+            // Env vars have the HIGHEST priority in LlmConfig::resolve_model()
+            // (env var > TOML > DB > default). If the .env file has e.g.
+            // NEARAI_MODEL=old-model, it shadows everything else. We must
+            // update this var or the /model change is invisible on restart.
+            let registry = crate::llm::ProviderRegistry::load();
+            let model_env = registry.model_env_var(&backend);
+            let env_var_prefix = format!("{}=", model_env);
+
+            // Only update the .env file if the var is actually set there
+            // (avoid injecting new vars the user never configured).
+            let env_path = crate::bootstrap::ironclaw_env_path();
+            let env_has_var = std::fs::read_to_string(&env_path)
+                .ok()
+                .is_some_and(|content| {
+                    content.lines().any(|line| {
+                        let trimmed = line.trim_start();
+                        !trimmed.starts_with('#') && trimmed.starts_with(&env_var_prefix)
+                    })
+                });
+            if env_has_var {
+                if let Err(e) = crate::bootstrap::upsert_bootstrap_var(model_env, &model_owned) {
+                    tracing::warn!("Failed to update {} in .env: {}", model_env, e);
+                } else {
+                    tracing::debug!("Updated {} in .env to {}", model_env, model_owned);
+                }
+            }
+
+            // 2b. Update (or create) the TOML config file.
+            //
+            // The TOML overlay has higher priority than DB settings on
+            // startup, so it MUST stay in sync with the DB.
             let toml_path = crate::settings::Settings::default_toml_path();
             match crate::settings::Settings::load_toml(&toml_path) {
                 Ok(Some(mut settings)) => {
@@ -856,7 +983,15 @@ impl Agent {
                     }
                 }
                 Ok(None) => {
-                    // No config file on disk; nothing to update.
+                    // No config file yet — create one so the model choice
+                    // survives restarts even when the DB is unavailable.
+                    let settings = crate::settings::Settings {
+                        selected_model: Some(model_owned),
+                        ..Default::default()
+                    };
+                    if let Err(e) = settings.save_toml(&toml_path) {
+                        tracing::warn!("Failed to create config.toml for model persistence: {}", e);
+                    }
                 }
                 Err(e) => {
                     tracing::warn!("Failed to load config.toml for model persistence: {}", e);
@@ -865,7 +1000,7 @@ impl Agent {
         })
         .await
         {
-            tracing::warn!("Model TOML persistence task failed: {}", e);
+            tracing::warn!("Model persistence task failed: {}", e);
         }
     }
 }

src/agent/dispatcher.rs

@@ -420,6 +420,19 @@ impl<'a> LoopDelegate for ChatDelegate<'a> {
         content: Option<String>,
         reason_ctx: &mut ReasoningContext,
     ) -> Result<Option<LoopOutcome>, Error> {
+        // Extract and sanitize the narrative before consuming `content`.
+        let narrative = content
+            .as_deref()
+            .filter(|c| !c.trim().is_empty())
+            .map(|c| {
+                let sanitized = self
+                    .agent
+                    .safety()
+                    .sanitize_tool_output("agent_narrative", c);
+                sanitized.content
+            })
+            .filter(|c| !c.trim().is_empty());
+
         // Add the assistant message with tool_calls to context.
         // OpenAI protocol requires this before tool-result messages.
         reason_ctx
@@ -440,6 +453,41 @@ impl<'a> LoopDelegate for ChatDelegate<'a> {
             )
             .await;
 
+        // Build per-tool decisions for the reasoning update.
+        // Sanitize each rationale through SafetyLayer (parity with JobDelegate).
+        let decisions: Vec<crate::channels::ToolDecision> = tool_calls
+            .iter()
+            .filter_map(|tc| {
+                tc.reasoning.as_ref().map(|r| {
+                    let sanitized = self
+                        .agent
+                        .safety()
+                        .sanitize_tool_output("tool_rationale", r)
+                        .content;
+                    crate::channels::ToolDecision {
+                        tool_name: tc.name.clone(),
+                        rationale: sanitized,
+                    }
+                })
+            })
+            .collect();
+
+        // Emit reasoning update to channels.
+        if narrative.is_some() || !decisions.is_empty() {
+            let _ = self
+                .agent
+                .channels
+                .send_status(
+                    &self.message.channel,
+                    StatusUpdate::ReasoningUpdate {
+                        narrative: narrative.clone().unwrap_or_default(),
+                        decisions: decisions.clone(),
+                    },
+                    &self.message.metadata,
+                )
+                .await;
+        }
+
         // Record tool calls in the thread with sensitive params redacted.
         {
             let mut redacted_args: Vec<serde_json::Value> = Vec::with_capacity(tool_calls.len());
@@ -455,8 +503,23 @@ impl<'a> LoopDelegate for ChatDelegate<'a> {
             if let Some(thread) = sess.threads.get_mut(&self.thread_id)
                 && let Some(turn) = thread.last_turn_mut()
             {
+                // Set turn-level narrative.
+                if turn.narrative.is_none() {
+                    turn.narrative = narrative;
+                }
                 for (tc, safe_args) in tool_calls.iter().zip(redacted_args) {
-                    turn.record_tool_call(&tc.name, safe_args);
+                    let sanitized_rationale = tc.reasoning.as_ref().map(|r| {
+                        self.agent
+                            .safety()
+                            .sanitize_tool_output("tool_rationale", r)
+                            .content
+                    });
+                    turn.record_tool_call_with_reasoning(
+                        &tc.name,
+                        safe_args,
+                        sanitized_rationale,
+                        Some(tc.id.clone()),
+                    );
                 }
             }
         }
@@ -726,7 +789,7 @@ impl<'a> LoopDelegate for ChatDelegate<'a> {
                         if let Some(thread) = sess.threads.get_mut(&self.thread_id)
                             && let Some(turn) = thread.last_turn_mut()
                         {
-                            turn.record_tool_error(error_msg.clone());
+                            turn.record_tool_error_for(&tc.id, error_msg.clone());
                         }
                     }
                     reason_ctx
@@ -852,16 +915,19 @@ impl<'a> LoopDelegate for ChatDelegate<'a> {
                         Err(e) => format!("Tool '{}' failed: {}", tc.name, e),
                     };
 
-                    // Record sanitized result in thread
+                    // Record sanitized result in thread (identity-based matching).
                     {
                         let mut sess = self.session.lock().await;
                         if let Some(thread) = sess.threads.get_mut(&self.thread_id)
                             && let Some(turn) = thread.last_turn_mut()
                         {
                             if is_tool_error {
-                                turn.record_tool_error(result_content.clone());
+                                turn.record_tool_error_for(&tc.id, result_content.clone());
                             } else {
-                                turn.record_tool_result(serde_json::json!(result_content));
+                                turn.record_tool_result_for(
+                                    &tc.id,
+                                    serde_json::json!(result_content),
+                                );
                             }
                         }
                     }
@@ -1098,15 +1164,23 @@ pub(crate) fn extract_suggestions(text: &str) -> (String, Vec<String>) {
         Regex::new(r"(?s)<suggestions>\s*(.*?)\s*</suggestions>").expect("valid regex") // safety: constant pattern
     });
 
-    // Find the position of the last closing code fence to avoid matching inside code blocks
+    // Build a sorted list of code fence positions to determine open/close pairing.
-    let last_code_fence = text.rfind("```").unwrap_or(0);
+    // A position is "inside" a fenced block when it falls between an odd-numbered
+    // fence (opening) and the next even-numbered fence (closing).
+    let fence_positions: Vec<usize> = text.match_indices("```").map(|(pos, _)| pos).collect();
 
-    // Find all matches, take the last one that's after the last code fence
+    let is_inside_fence = |pos: usize| -> bool {
+        // Count how many fences appear before `pos`. If odd, we're inside a fence.
+        let count = fence_positions.iter().take_while(|&&fp| fp <= pos).count();
+        count % 2 == 1
+    };
+
+    // Find all matches, take the last one that's outside any code fence
     let mut best_match: Option<regex::Match<'_>> = None;
     let mut best_capture: Option<String> = None;
     for caps in RE.captures_iter(text) {
         if let (Some(full), Some(inner)) = (caps.get(0), caps.get(1))
-            && full.start() >= last_code_fence
+            && !is_inside_fence(full.start())
         {
             best_match = Some(full);
             best_capture = Some(inner.as_str().to_string());
@@ -1225,6 +1299,7 @@ mod tests {
             document_extraction: None,
             sandbox_readiness: crate::agent::routine_engine::SandboxReadiness::DisabledByConfig,
             builder: None,
+            llm_backend: "nearai".to_string(),
         };
 
         Agent::new(
@@ -1453,11 +1528,13 @@ mod tests {
                     id: "call_2".to_string(),
                     name: "http".to_string(),
                     arguments: serde_json::json!({"url": "https://example.com"}),
+                    reasoning: None,
                 },
                 ToolCall {
                     id: "call_3".to_string(),
                     name: "echo".to_string(),
                     arguments: serde_json::json!({"message": "done"}),
+                    reasoning: None,
                 },
             ],
             user_timezone: None,
@@ -1643,6 +1720,7 @@ mod tests {
                     id: "call_1".to_string(),
                     name: "echo".to_string(),
                     arguments: serde_json::json!({"message": "hi"}),
+                    reasoning: None,
                 }],
             ),
             ChatMessage::tool_result("call_1", "echo", "hi"),
@@ -1735,11 +1813,13 @@ mod tests {
                         id: "c1".to_string(),
                         name: "http".to_string(),
                         arguments: serde_json::json!({}),
+                        reasoning: None,
                     },
                     ToolCall {
                         id: "c2".to_string(),
                         name: "echo".to_string(),
                         arguments: serde_json::json!({}),
+                        reasoning: None,
                     },
                 ],
             ),
@@ -1773,6 +1853,7 @@ mod tests {
                     id: "c1".to_string(),
                     name: "echo".to_string(),
                     arguments: serde_json::json!({}),
+                    reasoning: None,
                 }],
             ),
             ChatMessage::tool_result("c1", "echo", "done"),
@@ -1903,6 +1984,7 @@ mod tests {
                     id: crate::llm::generate_tool_call_id(0, 0),
                     name: "echo".to_string(),
                     arguments: serde_json::json!({"message": "looping"}),
+                    reasoning: None,
                 }],
                 input_tokens: 0,
                 output_tokens: 5,
@@ -2056,6 +2138,7 @@ mod tests {
                     id: crate::llm::generate_tool_call_id(0, 0),
                     name: "nonexistent_tool".to_string(),
                     arguments: serde_json::json!({}),
+                    reasoning: None,
                 }],
                 input_tokens: 0,
                 output_tokens: 5,
@@ -2092,6 +2175,7 @@ mod tests {
             document_extraction: None,
             sandbox_readiness: crate::agent::routine_engine::SandboxReadiness::DisabledByConfig,
             builder: None,
+            llm_backend: "nearai".to_string(),
         };
 
         Agent::new(
@@ -2212,6 +2296,7 @@ mod tests {
                 document_extraction: None,
                 sandbox_readiness: crate::agent::routine_engine::SandboxReadiness::DisabledByConfig,
                 builder: None,
+                llm_backend: "nearai".to_string(),
             };
 
             Agent::new(
@@ -2345,6 +2430,16 @@ mod tests {
         assert!(suggestions.is_empty()); // safety: test
     }
 
+    #[test]
+    fn test_extract_suggestions_inside_unclosed_code_fence() {
+        // Regression: odd number of fences (unclosed fence) must still be
+        // treated as "inside a code block".
+        let input = "```\ncode\n<suggestions>[\"bar\"]</suggestions>";
+        let (text, suggestions) = super::extract_suggestions(input);
+        assert_eq!(text, input); // safety: test
+        assert!(suggestions.is_empty()); // safety: test
+    }
+
     #[test]
     fn test_extract_suggestions_after_code_fence() {
         let input = "```\ncode\n```\nAnswer.\n<suggestions>[\"foo\"]</suggestions>";

src/agent/job_monitor.rs

@@ -21,8 +21,8 @@ use tokio::task::JoinHandle;
 use uuid::Uuid;
 
 use crate::channels::IncomingMessage;
-use crate::channels::web::types::SseEvent;
 use crate::context::{ContextManager, JobState};
+use ironclaw_common::AppEvent;
 
 /// Route context for forwarding job monitor events back to the user's channel.
 #[derive(Debug, Clone)]
@@ -36,15 +36,15 @@ pub struct JobMonitorRoute {
 /// injects assistant messages into the agent loop.
 ///
 /// The monitor forwards:
-/// - `SseEvent::JobMessage` (assistant role): injected as incoming messages so
+/// - `AppEvent::JobMessage` (assistant role): injected as incoming messages so
 ///   the main agent can read and relay to the user.
-/// - `SseEvent::JobResult`: injected as a completion notice, then the task exits.
+/// - `AppEvent::JobResult`: injected as a completion notice, then the task exits.
 ///
 /// Tool use/result and status events are intentionally skipped (too noisy for
 /// the main agent's context window).
 pub fn spawn_job_monitor(
     job_id: Uuid,
-    event_rx: broadcast::Receiver<(Uuid, SseEvent)>,
+    event_rx: broadcast::Receiver<(Uuid, String, AppEvent)>,
     inject_tx: mpsc::Sender<IncomingMessage>,
     route: JobMonitorRoute,
 ) -> JoinHandle<()> {
@@ -56,7 +56,7 @@ pub fn spawn_job_monitor(
 /// jobs don't stay `InProgress` forever in the `ContextManager`.
 pub fn spawn_job_monitor_with_context(
     job_id: Uuid,
-    mut event_rx: broadcast::Receiver<(Uuid, SseEvent)>,
+    mut event_rx: broadcast::Receiver<(Uuid, String, AppEvent)>,
     inject_tx: mpsc::Sender<IncomingMessage>,
     route: JobMonitorRoute,
     context_manager: Option<Arc<ContextManager>>,
@@ -68,13 +68,13 @@ pub fn spawn_job_monitor_with_context(
 
         loop {
             match event_rx.recv().await {
-                Ok((ev_job_id, event)) => {
+                Ok((ev_job_id, _user_id, event)) => {
                     if ev_job_id != job_id {
                         continue;
                     }
 
                     match event {
-                        SseEvent::JobMessage { role, content, .. } if role == "assistant" => {
+                        AppEvent::JobMessage { role, content, .. } if role == "assistant" => {
                             let mut msg = IncomingMessage::new(
                                 route.channel.clone(),
                                 route.user_id.clone(),
@@ -92,7 +92,7 @@ pub fn spawn_job_monitor_with_context(
                                 break;
                             }
                         }
-                        SseEvent::JobResult { status, .. } => {
+                        AppEvent::JobResult { status, .. } => {
                             // Transition in-memory state so the job frees its
                             // max_jobs slot and query tools show the final state.
                             if let Some(ref cm) = context_manager {
@@ -162,7 +162,7 @@ pub fn spawn_job_monitor_with_context(
 /// inject messages into) but we still need to free the `max_jobs` slot.
 pub fn spawn_completion_watcher(
     job_id: Uuid,
-    mut event_rx: broadcast::Receiver<(Uuid, SseEvent)>,
+    mut event_rx: broadcast::Receiver<(Uuid, String, AppEvent)>,
     context_manager: Arc<ContextManager>,
 ) -> JoinHandle<()> {
     let short_id = job_id.to_string()[..8].to_string();
@@ -170,7 +170,9 @@ pub fn spawn_completion_watcher(
     tokio::spawn(async move {
         loop {
             match event_rx.recv().await {
-                Ok((ev_job_id, SseEvent::JobResult { status, .. })) if ev_job_id == job_id => {
+                Ok((ev_job_id, _user_id, AppEvent::JobResult { status, .. }))
+                    if ev_job_id == job_id =>
+                {
                     let target = if status == "completed" {
                         JobState::Completed
                     } else {
@@ -227,7 +229,7 @@ mod tests {
 
     #[tokio::test]
     async fn test_monitor_forwards_assistant_messages() {
-        let (event_tx, _) = broadcast::channel::<(Uuid, SseEvent)>(16);
+        let (event_tx, _) = broadcast::channel::<(Uuid, String, AppEvent)>(16);
         let (inject_tx, mut inject_rx) = mpsc::channel::<IncomingMessage>(16);
 
         let job_id = Uuid::new_v4();
@@ -237,7 +239,8 @@ mod tests {
         event_tx
             .send((
                 job_id,
-                SseEvent::JobMessage {
+                "test-user".to_string(),
+                AppEvent::JobMessage {
                     job_id: job_id.to_string(),
                     role: "assistant".to_string(),
                     content: "I found a bug".to_string(),
@@ -259,7 +262,7 @@ mod tests {
 
     #[tokio::test]
     async fn test_monitor_ignores_other_jobs() {
-        let (event_tx, _) = broadcast::channel::<(Uuid, SseEvent)>(16);
+        let (event_tx, _) = broadcast::channel::<(Uuid, String, AppEvent)>(16);
         let (inject_tx, mut inject_rx) = mpsc::channel::<IncomingMessage>(16);
 
         let job_id = Uuid::new_v4();
@@ -270,7 +273,8 @@ mod tests {
         event_tx
             .send((
                 other_job_id,
-                SseEvent::JobMessage {
+                "test-user".to_string(),
+                AppEvent::JobMessage {
                     job_id: other_job_id.to_string(),
                     role: "assistant".to_string(),
                     content: "wrong job".to_string(),
@@ -289,7 +293,7 @@ mod tests {
 
     #[tokio::test]
     async fn test_monitor_exits_on_job_result() {
-        let (event_tx, _) = broadcast::channel::<(Uuid, SseEvent)>(16);
+        let (event_tx, _) = broadcast::channel::<(Uuid, String, AppEvent)>(16);
         let (inject_tx, mut inject_rx) = mpsc::channel::<IncomingMessage>(16);
 
         let job_id = Uuid::new_v4();
@@ -299,7 +303,8 @@ mod tests {
         event_tx
             .send((
                 job_id,
-                SseEvent::JobResult {
+                "test-user".to_string(),
+                AppEvent::JobResult {
                     job_id: job_id.to_string(),
                     status: "completed".to_string(),
                     session_id: None,
@@ -324,7 +329,7 @@ mod tests {
 
     #[tokio::test]
     async fn test_monitor_skips_tool_events() {
-        let (event_tx, _) = broadcast::channel::<(Uuid, SseEvent)>(16);
+        let (event_tx, _) = broadcast::channel::<(Uuid, String, AppEvent)>(16);
         let (inject_tx, mut inject_rx) = mpsc::channel::<IncomingMessage>(16);
 
         let job_id = Uuid::new_v4();
@@ -334,7 +339,8 @@ mod tests {
         event_tx
             .send((
                 job_id,
-                SseEvent::JobToolUse {
+                "test-user".to_string(),
+                AppEvent::JobToolUse {
                     job_id: job_id.to_string(),
                     tool_name: "shell".to_string(),
                     input: serde_json::json!({"command": "ls"}),
@@ -346,7 +352,8 @@ mod tests {
         event_tx
             .send((
                 job_id,
-                SseEvent::JobMessage {
+                "test-user".to_string(),
+                AppEvent::JobMessage {
                     job_id: job_id.to_string(),
                     role: "user".to_string(),
                     content: "user prompt".to_string(),
@@ -402,7 +409,7 @@ mod tests {
             .await
             .unwrap();
 
-        let (event_tx, _) = broadcast::channel::<(Uuid, SseEvent)>(16);
+        let (event_tx, _) = broadcast::channel::<(Uuid, String, AppEvent)>(16);
         let (inject_tx, mut inject_rx) = mpsc::channel::<IncomingMessage>(16);
 
         let handle = spawn_job_monitor_with_context(
@@ -417,7 +424,8 @@ mod tests {
         event_tx
             .send((
                 job_id,
-                SseEvent::JobResult {
+                "test-user".to_string(),
+                AppEvent::JobResult {
                     job_id: job_id.to_string(),
                     status: "completed".to_string(),
                     session_id: None,
@@ -450,7 +458,7 @@ mod tests {
             .await
             .unwrap();
 
-        let (event_tx, _) = broadcast::channel::<(Uuid, SseEvent)>(16);
+        let (event_tx, _) = broadcast::channel::<(Uuid, String, AppEvent)>(16);
         let (inject_tx, mut inject_rx) = mpsc::channel::<IncomingMessage>(16);
 
         let handle = spawn_job_monitor_with_context(
@@ -465,7 +473,8 @@ mod tests {
         event_tx
             .send((
                 job_id,
-                SseEvent::JobResult {
+                "test-user".to_string(),
+                AppEvent::JobResult {
                     job_id: job_id.to_string(),
                     status: "failed".to_string(),
                     session_id: None,
@@ -498,13 +507,14 @@ mod tests {
             .await
             .unwrap();
 
-        let (event_tx, _) = broadcast::channel::<(Uuid, SseEvent)>(16);
+        let (event_tx, _) = broadcast::channel::<(Uuid, String, AppEvent)>(16);
         let handle = spawn_completion_watcher(job_id, event_tx.subscribe(), Arc::clone(&cm));
 
         event_tx
             .send((
                 job_id,
-                SseEvent::JobResult {
+                "test-user".to_string(),
+                AppEvent::JobResult {
                     job_id: job_id.to_string(),
                     status: "completed".to_string(),
                     session_id: None,

src/agent/routine_engine.rs

@@ -18,13 +18,14 @@ use std::time::Duration;
 use chrono::Utc;
 use regex::Regex;
 use tokio::sync::{RwLock, mpsc};
+use tokio::task::JoinHandle;
 use uuid::Uuid;
 
 use crate::agent::Scheduler;
 use crate::agent::routine::{
     NotifyConfig, Routine, RoutineAction, RoutineRun, RunStatus, Trigger, next_cron_fire,
 };
-use crate::channels::OutgoingResponse;
+use crate::channels::{IncomingMessage, OutgoingResponse};
 use crate::config::RoutineConfig;
 use crate::context::{JobContext, JobState};
 use crate::db::Database;
@@ -45,6 +46,11 @@ enum EventMatcher {
     System { routine: Routine },
 }
 
+struct TriggeredRoutine {
+    routine: Routine,
+    detail: String,
+}
+
 /// Distinguishes why sandbox is unavailable so error messages are accurate.
 #[derive(Debug, Clone, Copy, PartialEq, Eq)]
 pub enum SandboxReadiness {
@@ -56,6 +62,40 @@ pub enum SandboxReadiness {
     DockerUnavailable,
 }
 
+/// Check whether an event-triggered routine's user/channel filters match an
+/// incoming message.
+///
+/// Returns `true` if:
+/// - The routine has an `Event` trigger (non-Event routines always return `false`)
+/// - The routine's `user_id` matches the message's user scope
+/// - The routine's channel filter (if any) matches the message channel
+///   case-insensitively
+///
+/// This is a pure function extracted from `check_event_triggers` so the
+/// filter logic can be unit-tested without async infrastructure.
+pub(crate) fn routine_matches_message(routine: &Routine, message: &IncomingMessage) -> bool {
+    // Only Event-triggered routines can match incoming messages.
+    if !matches!(routine.trigger, Trigger::Event { .. }) {
+        return false;
+    }
+
+    // User ownership filter — only fire routines scoped to this user.
+    if routine.user_id != message.user_id {
+        return false;
+    }
+
+    // Channel filter (case-insensitive, matching emit_system_event behavior)
+    if let Trigger::Event {
+        channel: Some(ch), ..
+    } = &routine.trigger
+        && !ch.eq_ignore_ascii_case(&message.channel)
+    {
+        return false;
+    }
+
+    true
+}
+
 /// The routine execution engine.
 pub struct RoutineEngine {
     config: RoutineConfig,
@@ -167,10 +207,45 @@ impl RoutineEngine {
     }
 
     /// Check incoming message against event triggers. Returns number of routines fired.
+    pub async fn check_event_triggers(&self, message: &IncomingMessage, content: &str) -> usize {
+        let triggered = self.matching_event_triggers(message, content).await;
+        let fired = triggered.len();
+        for triggered in triggered {
+            std::mem::drop(self.spawn_fire(triggered.routine, "event", Some(triggered.detail)));
+        }
+        fired
+    }
+
+    /// Fire matching event-triggered routines and wait for them to complete.
     ///
-    /// Accepts only the three fields needed for matching (user scope, channel,
+    /// Used by single-message REPL mode so the process does not exit before
-    /// message content) so callers never need to clone a full `IncomingMessage`.
+    /// background event-triggered routines finish.
-    pub async fn check_event_triggers(&self, user_id: &str, channel: &str, content: &str) -> usize {
+    pub async fn check_event_triggers_and_wait(
+        &self,
+        message: &IncomingMessage,
+        content: &str,
+    ) -> usize {
+        let triggered = self.matching_event_triggers(message, content).await;
+        let fired = triggered.len();
+        let handles: Vec<JoinHandle<()>> = triggered
+            .into_iter()
+            .map(|triggered| self.spawn_fire(triggered.routine, "event", Some(triggered.detail)))
+            .collect();
+
+        for handle in handles {
+            if let Err(e) = handle.await {
+                tracing::warn!(error = %e, "Event-triggered routine task failed");
+            }
+        }
+
+        fired
+    }
+
+    async fn matching_event_triggers(
+        &self,
+        message: &IncomingMessage,
+        content: &str,
+    ) -> Vec<TriggeredRoutine> {
         let cache = self.event_cache.read().await;
 
         // Early return if there are no message matchers at all.
@@ -178,10 +253,9 @@ impl RoutineEngine {
             .iter()
             .any(|m| matches!(m, EventMatcher::Message { .. }))
         {
-            return 0;
+            return Vec::new();
         }
-
+        let mut triggered = Vec::new();
-        let mut fired = 0;
 
         // Collect routine IDs for batch query
         let routine_ids: Vec<Uuid> = cache
@@ -193,13 +267,13 @@ impl RoutineEngine {
             .collect();
 
         if routine_ids.is_empty() {
-            return 0;
+            return Vec::new();
         }
 
         // Single batch query instead of N queries
         let concurrent_counts = match self.batch_concurrent_counts(&routine_ids).await {
             Some(counts) => counts,
-            None => return 0,
+            None => return Vec::new(),
         };
 
         for matcher in cache.iter() {
@@ -208,16 +282,24 @@ impl RoutineEngine {
                 EventMatcher::System { .. } => continue,
             };
 
-            if routine.user_id != user_id {
+            // User ownership + channel filter (extracted for testability).
-                continue;
+            if !routine_matches_message(routine, message) {
-            }
+                // User mismatch is expected for multi-user setups — keep at
-
+                // trace to avoid one log per routine per inbound message.
-            // Channel filter
+                if routine.user_id != message.user_id {
-            if let Trigger::Event {
+                    tracing::trace!(
-                channel: Some(ch), ..
+                        routine = %routine.name,
-            } = &routine.trigger
+                        routine_user = %routine.user_id,
-                && ch != channel
+                        message_user = %message.user_id,
-            {
+                        "Skipped: user scope mismatch"
+                    );
+                } else {
+                    tracing::debug!(
+                        routine = %routine.name,
+                        channel = %message.channel,
+                        "Skipped: channel mismatch"
+                    );
+                }
                 continue;
             }
 
@@ -228,14 +310,14 @@ impl RoutineEngine {
 
             // Cooldown check
             if !self.check_cooldown(routine) {
-                tracing::trace!(routine = %routine.name, "Skipped: cooldown active");
+                tracing::debug!(routine = %routine.name, "Skipped: cooldown active");
                 continue;
             }
 
             // Concurrent run check (using batch-loaded counts)
             let running_count = concurrent_counts.get(&routine.id).copied().unwrap_or(0);
             if running_count >= routine.guardrails.max_concurrent as i64 {
-                tracing::trace!(routine = %routine.name, "Skipped: max concurrent reached");
+                tracing::debug!(routine = %routine.name, "Skipped: max concurrent reached");
                 continue;
             }
 
@@ -246,11 +328,13 @@ impl RoutineEngine {
             }
 
             let detail = truncate(content, 200);
-            self.spawn_fire(routine.clone(), "event", Some(detail));
+            triggered.push(TriggeredRoutine {
-            fired += 1;
+                routine: routine.clone(),
+                detail,
+            });
         }
 
-        fired
+        triggered
     }
 
     /// Emit a structured event to system-event routines.
@@ -806,7 +890,12 @@ impl RoutineEngine {
     }
 
     /// Spawn a fire in a background task.
-    fn spawn_fire(&self, routine: Routine, trigger_type: &str, trigger_detail: Option<String>) {
+    fn spawn_fire(
+        &self,
+        routine: Routine,
+        trigger_type: &str,
+        trigger_detail: Option<String>,
+    ) -> JoinHandle<()> {
         let run = RoutineRun {
             id: Uuid::new_v4(),
             routine_id: routine.id,
@@ -843,7 +932,7 @@ impl RoutineEngine {
                 return;
             }
             execute_routine(engine, routine, run).await;
-        });
+        })
     }
 
     fn check_cooldown(&self, routine: &Routine) -> bool {
@@ -1305,6 +1394,19 @@ async fn execute_lightweight(
     }
 }
 
+/// Sanitize a user-controlled string before interpolation into an LLM prompt.
+/// Strips newlines (which could break prompt structure) and truncates to a
+/// reasonable length to limit abuse surface.
+fn sanitize_prompt_field(value: &str) -> String {
+    const MAX_LEN: usize = 128;
+    value
+        .chars()
+        .filter(|&c| c != '\n' && c != '\r')
+        .take(MAX_LEN)
+        .map(|c| if c == '`' { '\'' } else { c })
+        .collect()
+}
+
 fn build_lightweight_prompt(
     prompt: &str,
     context_parts: &[String],
@@ -1323,14 +1425,16 @@ fn build_lightweight_prompt(
         );
 
         if let Some(channel) = notify.channel.as_deref() {
+            let sanitized = sanitize_prompt_field(channel);
             full_prompt.push_str(&format!(
-                "The configured delivery channel for this routine is `{channel}`.\n"
+                "The configured delivery channel for this routine is `{sanitized}`.\n"
             ));
         }
 
         if let Some(user) = notify.user.as_deref() {
+            let sanitized = sanitize_prompt_field(user);
             full_prompt.push_str(&format!(
-                "The configured delivery target for this routine is `{user}`.\n"
+                "The configured delivery target for this routine is `{sanitized}`.\n"
             ));
         }
 
@@ -1440,6 +1544,7 @@ fn handle_text_response(
 /// This is a simplified version of the full dispatcher loop:
 /// - Max 3-5 iterations (configurable)
 /// - Sequential tool execution (not parallel)
+/// - Uses the owner's live autonomous tool scope when lightweight tools are enabled
 /// - Auto-approval of non-Always tools
 /// - No hooks or approval dialogs
 async fn execute_lightweight_with_tools(
@@ -1486,7 +1591,10 @@ async fn execute_lightweight_with_tools(
         let force_text = iteration >= max_iterations;
 
         if force_text {
-            // Final iteration: no tools, just get text response
+            // Final iteration: no tools, just get text response.
+            // Claude 4.6 rejects assistant prefill; NEAR AI rejects any non-user-ending
+            // conversation. Ensure the last message is user-role.
+            crate::util::ensure_ends_with_user_message(&mut messages);
             let request = CompletionRequest::new(messages)
                 .with_max_tokens(effective_max_tokens)
                 .with_temperature(0.3);
@@ -1765,6 +1873,13 @@ pub fn spawn_cron_ticker(
         engine.check_cron_triggers().await;
 
         let mut ticker = tokio::time::interval(interval);
+        ticker.set_missed_tick_behavior(tokio::time::MissedTickBehavior::Skip);
+        // Periodic event cache refresh so web/CLI mutations are picked up
+        // without requiring tool-path code to call refresh_event_cache().
+        // Uses wall-clock elapsed time so the refresh cadence is stable
+        // regardless of the cron tick interval configuration.
+        let refresh_interval = Duration::from_secs(60);
+        let mut last_refresh = tokio::time::Instant::now();
 
         loop {
             ticker.tick().await;
@@ -1772,7 +1887,11 @@ pub fn spawn_cron_ticker(
             // never races with FullJobWatcher instances from this process.
             engine.sync_dispatched_runs().await;
             engine.check_cron_triggers().await;
-            engine.sync_dispatched_runs().await;
+
+            if last_refresh.elapsed() >= refresh_interval {
+                engine.refresh_event_cache().await;
+                last_refresh = tokio::time::Instant::now();
+            }
         }
     })
 }
@@ -1838,7 +1957,13 @@ fn strip_html_tags(s: &str) -> String {
 
 #[cfg(test)]
 mod tests {
-    use crate::agent::routine::{NotifyConfig, RunStatus};
+    use chrono::Utc;
+    use uuid::Uuid;
+
+    use crate::agent::routine::{
+        NotifyConfig, Routine, RoutineAction, RoutineGuardrails, RunStatus, Trigger,
+    };
+    use crate::channels::IncomingMessage;
     use crate::config::RoutineConfig;
 
     #[test]
@@ -2036,6 +2161,117 @@ mod tests {
         }
     }
 
+    /// Helper to build a test routine with the given user_id and trigger.
+    fn make_routine(user_id: &str, trigger: Trigger) -> Routine {
+        Routine {
+            id: Uuid::new_v4(),
+            name: "test".to_string(),
+            description: String::new(),
+            user_id: user_id.to_string(),
+            enabled: true,
+            trigger,
+            action: RoutineAction::Lightweight {
+                prompt: String::new(),
+                context_paths: vec![],
+                max_tokens: 1000,
+                use_tools: false,
+                max_tool_rounds: 0,
+            },
+            guardrails: RoutineGuardrails::default(),
+            notify: Default::default(),
+            last_run_at: None,
+            next_fire_at: None,
+            run_count: 0,
+            consecutive_failures: 0,
+            state: serde_json::Value::Null,
+            created_at: Utc::now(),
+            updated_at: Utc::now(),
+        }
+    }
+
+    /// Helper to build a test IncomingMessage.
+    fn make_message(user_id: &str, channel: &str, content: &str) -> IncomingMessage {
+        IncomingMessage {
+            id: Uuid::new_v4(),
+            channel: channel.to_string(),
+            user_id: user_id.to_string(),
+            owner_id: user_id.to_string(),
+            sender_id: user_id.to_string(),
+            user_name: None,
+            content: content.to_string(),
+            thread_id: None,
+            conversation_scope_id: None,
+            received_at: Utc::now(),
+            metadata: serde_json::Value::Null,
+            timezone: None,
+            attachments: vec![],
+            is_internal: false,
+        }
+    }
+
+    /// Regression test for issue #1051: event triggers used case-sensitive
+    /// channel comparison, so "Telegram" != "telegram" caused silent mismatch.
+    /// Tests the actual `routine_matches_message` function used in `check_event_triggers`.
+    #[test]
+    fn test_channel_filter_is_case_insensitive() {
+        let routine = make_routine(
+            "user1",
+            Trigger::Event {
+                pattern: ".*".to_string(),
+                channel: Some("Telegram".to_string()),
+            },
+        );
+        let msg = make_message("user1", "telegram", "hello");
+
+        // Case-insensitive channel match must succeed
+        assert!(super::routine_matches_message(&routine, &msg));
+
+        // Exact case must also work
+        let msg_exact = make_message("user1", "Telegram", "hello");
+        assert!(super::routine_matches_message(&routine, &msg_exact));
+
+        // Different channel must not match
+        let msg_wrong = make_message("user1", "discord", "hello");
+        assert!(!super::routine_matches_message(&routine, &msg_wrong));
+    }
+
+    /// Regression test for issue #1051: event triggers did not filter by
+    /// user_id, so routines from user A could fire on messages from user B.
+    /// Tests the actual `routine_matches_message` function used in `check_event_triggers`.
+    #[test]
+    fn test_event_trigger_requires_user_match() {
+        let routine = make_routine(
+            "alice",
+            Trigger::Event {
+                pattern: ".*".to_string(),
+                channel: None,
+            },
+        );
+
+        // Different user must not match
+        let msg_bob = make_message("bob", "telegram", "hello");
+        assert!(!super::routine_matches_message(&routine, &msg_bob));
+
+        // Same user must match
+        let msg_alice = make_message("alice", "telegram", "hello");
+        assert!(super::routine_matches_message(&routine, &msg_alice));
+    }
+
+    /// When no channel filter is set, any channel should match (given user matches).
+    #[test]
+    fn test_no_channel_filter_matches_any_channel() {
+        let routine = make_routine(
+            "user1",
+            Trigger::Event {
+                pattern: ".*".to_string(),
+                channel: None,
+            },
+        );
+
+        let msg = make_message("user1", "whatever_channel", "hello");
+        assert!(super::routine_matches_message(&routine, &msg));
+    }
+
     #[test]
     fn test_routine_tool_denylist_blocks_self_management_tools() {
         let denylisted = vec![

src/agent/scheduler.rs

@@ -9,7 +9,6 @@ use tokio::task::JoinHandle;
 use uuid::Uuid;
 
 use crate::agent::task::{Task, TaskContext, TaskOutput};
-use crate::channels::web::types::SseEvent;
 use crate::config::AgentConfig;
 use crate::context::{ContextManager, JobContext, JobState};
 use crate::db::Database;
@@ -67,8 +66,8 @@ pub struct Scheduler {
     extension_manager: Option<Arc<ExtensionManager>>,
     store: Option<Arc<dyn Database>>,
     hooks: Arc<HookRegistry>,
-    /// SSE broadcast sender for live job event streaming.
+    /// SSE manager for live job event streaming.
-    sse_tx: Option<tokio::sync::broadcast::Sender<SseEvent>>,
+    sse_tx: Option<Arc<crate::channels::web::sse::SseManager>>,
     /// HTTP interceptor for trace recording/replay (propagated to workers).
     http_interceptor: Option<Arc<dyn crate::llm::recording::HttpInterceptor>>,
     /// Running jobs (main LLM-driven jobs).
@@ -102,9 +101,9 @@ impl Scheduler {
         }
     }
 
-    /// Set the SSE broadcast sender for live job event streaming.
+    /// Set the SSE manager for live job event streaming.
-    pub fn set_sse_sender(&mut self, tx: tokio::sync::broadcast::Sender<SseEvent>) {
+    pub fn set_sse_sender(&mut self, sse: Arc<crate::channels::web::sse::SseManager>) {
-        self.sse_tx = Some(tx);
+        self.sse_tx = Some(sse);
     }
 
     /// Set the HTTP interceptor for trace recording/replay.

src/agent/session.rs

@@ -16,8 +16,8 @@ use chrono::{DateTime, TimeDelta, Utc};
 use serde::{Deserialize, Serialize};
 use uuid::Uuid;
 
-use crate::channels::web::util::truncate_preview;
 use crate::llm::{ChatMessage, ToolCall, generate_tool_call_id};
+use ironclaw_common::truncate_preview;
 
 /// A session containing one or more threads.
 #[derive(Debug, Clone, Serialize, Deserialize)]
@@ -449,6 +449,7 @@ impl Thread {
                         id: call_id.clone(),
                         name: tc.name.clone(),
                         arguments: tc.parameters.clone(),
+                        reasoning: None,
                     })
                     .collect();
 
@@ -522,7 +523,12 @@ impl Thread {
                             && let Some(ref tcs) = assistant_msg.tool_calls
                         {
                             for tc in tcs {
-                                turn.record_tool_call(&tc.name, tc.arguments.clone());
+                                turn.record_tool_call_with_reasoning(
+                                    &tc.name,
+                                    tc.arguments.clone(),
+                                    tc.reasoning.clone(),
+                                    Some(tc.id.clone()),
+                                );
                             }
                         }
 
@@ -602,6 +608,10 @@ pub struct Turn {
     pub completed_at: Option<DateTime<Utc>>,
     /// Error message (if failed).
     pub error: Option<String>,
+    /// Agent's reasoning narrative for this turn.
+    /// Cleaned via `clean_response` and sanitized through `SafetyLayer` before storage.
+    #[serde(default, skip_serializing_if = "Option::is_none")]
+    pub narrative: Option<String>,
     /// Transient image content parts for multimodal LLM input.
     /// Not serialized — images are only needed for the current LLM call.
     /// The text description in `user_input` persists for compaction/context.
@@ -621,6 +631,7 @@ impl Turn {
             started_at: Utc::now(),
             completed_at: None,
             error: None,
+            narrative: None,
             image_content_parts: Vec::new(),
         }
     }
@@ -656,6 +667,26 @@ impl Turn {
             parameters: params,
             result: None,
             error: None,
+            rationale: None,
+            tool_call_id: None,
+        });
+    }
+
+    /// Record a tool call with reasoning context.
+    pub fn record_tool_call_with_reasoning(
+        &mut self,
+        name: impl Into<String>,
+        params: serde_json::Value,
+        rationale: Option<String>,
+        tool_call_id: Option<String>,
+    ) {
+        self.tool_calls.push(TurnToolCall {
+            name: name.into(),
+            parameters: params,
+            result: None,
+            error: None,
+            rationale,
+            tool_call_id,
         });
     }
 
@@ -672,6 +703,60 @@ impl Turn {
             call.error = Some(error.into());
         }
     }
+
+    /// Record a tool result by tool_call_id, with fallback to first pending call.
+    pub fn record_tool_result_for(&mut self, tool_call_id: &str, result: serde_json::Value) {
+        if let Some(call) = self
+            .tool_calls
+            .iter_mut()
+            .find(|c| c.tool_call_id.as_deref() == Some(tool_call_id))
+        {
+            call.result = Some(result);
+        } else if let Some(call) = self
+            .tool_calls
+            .iter_mut()
+            .find(|c| c.result.is_none() && c.error.is_none())
+        {
+            tracing::debug!(
+                tool_call_id = %tool_call_id,
+                fallback_tool = %call.name,
+                "tool_call_id not found, falling back to first pending call"
+            );
+            call.result = Some(result);
+        } else {
+            tracing::warn!(
+                tool_call_id = %tool_call_id,
+                "Tool result dropped: no matching or pending tool call"
+            );
+        }
+    }
+
+    /// Record a tool error by tool_call_id, with fallback to first pending call.
+    pub fn record_tool_error_for(&mut self, tool_call_id: &str, error: impl Into<String>) {
+        if let Some(call) = self
+            .tool_calls
+            .iter_mut()
+            .find(|c| c.tool_call_id.as_deref() == Some(tool_call_id))
+        {
+            call.error = Some(error.into());
+        } else if let Some(call) = self
+            .tool_calls
+            .iter_mut()
+            .find(|c| c.result.is_none() && c.error.is_none())
+        {
+            tracing::debug!(
+                tool_call_id = %tool_call_id,
+                fallback_tool = %call.name,
+                "tool_call_id not found, falling back to first pending call"
+            );
+            call.error = Some(error.into());
+        } else {
+            tracing::warn!(
+                tool_call_id = %tool_call_id,
+                "Tool error dropped: no matching or pending tool call"
+            );
+        }
+    }
 }
 
 /// Record of a tool call made during a turn.
@@ -685,6 +770,12 @@ pub struct TurnToolCall {
     pub result: Option<serde_json::Value>,
     /// Error from the tool (if failed).
     pub error: Option<String>,
+    /// Agent's reasoning for choosing this tool.
+    #[serde(default, skip_serializing_if = "Option::is_none")]
+    pub rationale: Option<String>,
+    /// The tool_call_id from the LLM, for identity-based result matching.
+    #[serde(default, skip_serializing_if = "Option::is_none")]
+    pub tool_call_id: Option<String>,
 }
 
 #[cfg(test)]
@@ -1309,6 +1400,7 @@ mod tests {
             id: "call_0".to_string(),
             name: "search".to_string(),
             arguments: serde_json::json!({"q": "test"}),
+            reasoning: None,
         };
         let messages = vec![
             ChatMessage::user("Find test"),
@@ -1339,6 +1431,7 @@ mod tests {
             id: "call_0".to_string(),
             name: "http".to_string(),
             arguments: serde_json::json!({}),
+            reasoning: None,
         };
         let messages = vec![
             ChatMessage::user("Fetch URL"),
@@ -1404,11 +1497,13 @@ mod tests {
             id: "call_a".to_string(),
             name: "search".to_string(),
             arguments: serde_json::json!({"q": "data"}),
+            reasoning: None,
         };
         let tc2 = ToolCall {
             id: "call_b".to_string(),
             name: "write".to_string(),
             arguments: serde_json::json!({"path": "out.txt"}),
+            reasoning: None,
         };
         let messages = vec![
             ChatMessage::user("Find and save"),
@@ -1620,4 +1715,100 @@ mod tests {
         let merged = thread.drain_pending_messages().unwrap();
         assert_eq!(merged, "failed batch\nnew msg");
     }
+
+    #[test]
+    fn test_record_tool_result_for_by_id() {
+        let mut turn = Turn::new(0, "test");
+        turn.record_tool_call_with_reasoning(
+            "tool_a",
+            serde_json::json!({}),
+            None,
+            Some("id_a".into()),
+        );
+        turn.record_tool_call_with_reasoning(
+            "tool_b",
+            serde_json::json!({}),
+            None,
+            Some("id_b".into()),
+        );
+
+        // Record result for second tool by ID
+        turn.record_tool_result_for("id_b", serde_json::json!("result_b"));
+        assert!(turn.tool_calls[0].result.is_none());
+        assert_eq!(
+            turn.tool_calls[1].result.as_ref().unwrap(),
+            &serde_json::json!("result_b")
+        );
+    }
+
+    #[test]
+    fn test_record_tool_error_for_by_id() {
+        let mut turn = Turn::new(0, "test");
+        turn.record_tool_call_with_reasoning(
+            "tool_a",
+            serde_json::json!({}),
+            None,
+            Some("id_a".into()),
+        );
+        turn.record_tool_call_with_reasoning(
+            "tool_b",
+            serde_json::json!({}),
+            None,
+            Some("id_b".into()),
+        );
+
+        turn.record_tool_error_for("id_a", "failed");
+        assert_eq!(turn.tool_calls[0].error.as_deref(), Some("failed"));
+        assert!(turn.tool_calls[1].error.is_none());
+    }
+
+    #[test]
+    fn test_record_tool_result_for_fallback_to_pending() {
+        let mut turn = Turn::new(0, "test");
+        turn.record_tool_call_with_reasoning(
+            "tool_a",
+            serde_json::json!({}),
+            None,
+            Some("id_a".into()),
+        );
+        turn.record_tool_call_with_reasoning(
+            "tool_b",
+            serde_json::json!({}),
+            None,
+            Some("id_b".into()),
+        );
+
+        // First tool already has a result
+        turn.tool_calls[0].result = Some(serde_json::json!("done"));
+
+        // Unknown ID should fall back to first pending (tool_b)
+        turn.record_tool_result_for("unknown_id", serde_json::json!("fallback"));
+        assert_eq!(
+            turn.tool_calls[0].result.as_ref().unwrap(),
+            &serde_json::json!("done")
+        );
+        assert_eq!(
+            turn.tool_calls[1].result.as_ref().unwrap(),
+            &serde_json::json!("fallback")
+        );
+    }
+
+    #[test]
+    fn test_record_tool_result_for_no_pending_is_noop() {
+        let mut turn = Turn::new(0, "test");
+        turn.record_tool_call_with_reasoning(
+            "tool_a",
+            serde_json::json!({}),
+            None,
+            Some("id_a".into()),
+        );
+        turn.tool_calls[0].result = Some(serde_json::json!("done"));
+
+        // No pending calls, unknown ID — should be a no-op
+        turn.record_tool_result_for("unknown_id", serde_json::json!("lost"));
+        assert_eq!(
+            turn.tool_calls[0].result.as_ref().unwrap(),
+            &serde_json::json!("done")
+        );
+    }
 }

src/agent/session_manager.rs

@@ -102,11 +102,30 @@ impl SessionManager {
     /// Resolve an external thread ID to an internal thread.
     ///
     /// Returns the session and thread ID. Creates both if they don't exist.
+    /// Delegates to [`resolve_thread_with_parsed_uuid`](Self::resolve_thread_with_parsed_uuid)
+    /// with `parsed_uuid: None`.
     pub async fn resolve_thread(
         &self,
         user_id: &str,
         channel: &str,
         external_thread_id: Option<&str>,
+    ) -> (Arc<Mutex<Session>>, Uuid) {
+        self.resolve_thread_with_parsed_uuid(user_id, channel, external_thread_id, None)
+            .await
+    }
+
+    /// Like [`resolve_thread`](Self::resolve_thread), but accepts a pre-parsed
+    /// UUID to skip redundant parsing when the caller has already validated
+    /// the external thread ID as a UUID (e.g. the approval routing path).
+    ///
+    /// Uses a single read-lock acquisition for both the key lookup and the UUID
+    /// adoption check to reduce contention under concurrent approval load.
+    pub async fn resolve_thread_with_parsed_uuid(
+        &self,
+        user_id: &str,
+        channel: &str,
+        external_thread_id: Option<&str>,
+        parsed_uuid: Option<Uuid>,
     ) -> (Arc<Mutex<Session>>, Uuid) {
         let session = self.get_or_create_session(user_id).await;
 
@@ -116,51 +135,65 @@ impl SessionManager {
             external_thread_id: external_thread_id.map(String::from),
         };
 
-        // Check if we have a mapping
+        // Use pre-parsed UUID if available, otherwise parse from string.
-        {
+        let ext_uuid = parsed_uuid
+            .or_else(|| external_thread_id.and_then(|ext_tid| Uuid::parse_str(ext_tid).ok()));
+
+        // Validate that parsed_uuid (if provided) is consistent with external_thread_id.
+        #[cfg(debug_assertions)]
+        if let (Some(parsed), Some(ext_tid)) = (&parsed_uuid, external_thread_id) {
+            debug_assert_eq!(
+                Uuid::parse_str(ext_tid).ok().as_ref(),
+                Some(parsed),
+                "parsed_uuid must be the parsed form of external_thread_id"
+            );
+        }
+
+        // Single read lock for both the key lookup and UUID adoption check
+        let adoptable_uuid = {
             let thread_map = self.thread_map.read().await;
+
+            // Fast path: exact key match
             if let Some(&thread_id) = thread_map.get(&key) {
-                // Verify thread still exists in session
                 let sess = session.lock().await;
                 if sess.threads.contains_key(&thread_id) {
                     return (Arc::clone(&session), thread_id);
                 }
             }
-        }
 
-        // Check if external_thread_id is itself a known thread UUID that
+            // UUID adoption check (still under the same read lock).
-        // exists in the session but was never registered in the thread_map
+            // If external_thread_id is a valid UUID not mapped elsewhere,
-        // (e.g. created by chat_new_thread_handler or hydrated from DB).
+            // it may be a thread created by chat_new_thread_handler or
-        // We only adopt it if no thread_map entry maps to this UUID —
+            // hydrated from DB that we can adopt.
-        // otherwise it belongs to a different channel scope.
+            // Only attempt adoption when external_thread_id is Some, preserving
-        if let Some(ext_tid) = external_thread_id
+            // the invariant that None external_thread_id never triggers adoption.
-            && let Ok(ext_uuid) = Uuid::parse_str(ext_tid)
+            if external_thread_id.is_some() {
-        {
+                ext_uuid.filter(|&uuid| !thread_map.values().any(|&v| v == uuid))
-            let thread_map = self.thread_map.read().await;
+            } else {
-            let mapped_elsewhere = thread_map.values().any(|&v| v == ext_uuid);
+                None
-            drop(thread_map);
+            }
+        }; // Single read lock dropped here
 
-            if !mapped_elsewhere {
+        // If we found an adoptable UUID, verify it exists in session and acquire write lock
-                let sess = session.lock().await;
+        if let Some(ext_uuid) = adoptable_uuid {
-                if sess.threads.contains_key(&ext_uuid) {
+            let sess = session.lock().await;
-                    drop(sess);
+            if sess.threads.contains_key(&ext_uuid) {
+                drop(sess);
 
-                    let mut thread_map = self.thread_map.write().await;
+                let mut thread_map = self.thread_map.write().await;
-                    // Re-check after acquiring write lock to prevent race condition
+                // Re-check after acquiring write lock to prevent race condition
-                    // where another task mapped this UUID between our read and write.
+                // where another task mapped this UUID between our read and write.
-                    if !thread_map.values().any(|&v| v == ext_uuid) {
+                if !thread_map.values().any(|&v| v == ext_uuid) {
-                        thread_map.insert(key, ext_uuid);
+                    thread_map.insert(key, ext_uuid);
-                        drop(thread_map);
+                    drop(thread_map);
-                        // Ensure undo manager exists
+                    // Ensure undo manager exists
-                        let mut undo_managers = self.undo_managers.write().await;
+                    let mut undo_managers = self.undo_managers.write().await;
-                        undo_managers
+                    undo_managers
-                            .entry(ext_uuid)
+                        .entry(ext_uuid)
-                            .or_insert_with(|| Arc::new(Mutex::new(UndoManager::new())));
+                        .or_insert_with(|| Arc::new(Mutex::new(UndoManager::new())));
-                        return (session, ext_uuid);
+                    return (session, ext_uuid);
-                    }
-                    // If it was mapped elsewhere while we were unlocked, fall through
-                    // to create a new thread, preserving channel isolation.
                 }
+                // If mapped elsewhere while unlocked, fall through to create new thread
             }
         }
 
@@ -909,6 +942,44 @@ mod tests {
         }
     }
 
+    #[tokio::test]
+    async fn test_resolve_thread_consolidates_read_path() {
+        // Verify that resolve_thread still correctly handles:
+        // 1. Fast path: key exists in thread_map
+        // 2. UUID adoption: external_thread_id is a UUID in session but not in map
+        // 3. New thread: neither path matches
+        use crate::agent::session::Thread;
+
+        let manager = SessionManager::new();
+
+        // Case 1: Normal resolution creates thread and maps it
+        let (session1, tid1) = manager
+            .resolve_thread("user1", "chan1", Some("ext-1"))
+            .await;
+        // Resolving again with same key should return same thread (fast path)
+        let (_, tid1_again) = manager
+            .resolve_thread("user1", "chan1", Some("ext-1"))
+            .await;
+        assert_eq!(tid1, tid1_again);
+
+        // Case 2: UUID adoption - insert a thread directly into session
+        let adopted_id = Uuid::new_v4();
+        {
+            let mut sess = session1.lock().await;
+            let thread = Thread::with_id(adopted_id, sess.id);
+            sess.threads.insert(adopted_id, thread);
+        }
+        // Resolve with the UUID as external_thread_id -- should adopt it
+        let (_, resolved) = manager
+            .resolve_thread("user1", "chan1", Some(&adopted_id.to_string()))
+            .await;
+        assert_eq!(resolved, adopted_id);
+
+        // Case 3: Different channel gets different thread
+        let (_, tid2) = manager.resolve_thread("user1", "chan2", None).await;
+        assert_ne!(tid1, tid2);
+    }
+
     #[tokio::test]
     async fn test_resolve_thread_finds_existing_session_thread_by_uuid() {
         use crate::agent::session::{Session, Thread};
@@ -947,4 +1018,88 @@ mod tests {
             "should have exactly 1 thread, not a duplicate"
         );
     }
+
+    #[tokio::test]
+    async fn test_resolve_thread_with_pre_parsed_uuid_adopts_thread() {
+        use crate::agent::session::Thread;
+
+        let manager = SessionManager::new();
+        let (session, _) = manager.resolve_thread("user1", "chan1", None).await;
+
+        // Manually insert a thread with a known UUID
+        let known_id = Uuid::new_v4();
+        {
+            let mut sess = session.lock().await;
+            let thread = Thread::with_id(known_id, sess.id);
+            sess.threads.insert(known_id, thread);
+        }
+
+        // Resolve with pre-parsed UUID -- should adopt it without re-parsing
+        let (_, resolved) = manager
+            .resolve_thread_with_parsed_uuid(
+                "user1",
+                "chan1",
+                Some(&known_id.to_string()),
+                Some(known_id),
+            )
+            .await;
+        assert_eq!(resolved, known_id);
+    }
+
+    #[tokio::test]
+    async fn test_resolve_thread_with_parsed_uuid_none_delegates_to_parse() {
+        use crate::agent::session::Thread;
+
+        let manager = SessionManager::new();
+        let (session, _) = manager.resolve_thread("user2", "chan2", None).await;
+
+        // Insert a thread with a known UUID
+        let known_id = Uuid::new_v4();
+        {
+            let mut sess = session.lock().await;
+            let thread = Thread::with_id(known_id, sess.id);
+            sess.threads.insert(known_id, thread);
+        }
+
+        // Resolve with parsed_uuid=None but a valid UUID string -- should
+        // fall back to parsing the string and still adopt the thread
+        let (_, resolved) = manager
+            .resolve_thread_with_parsed_uuid("user2", "chan2", Some(&known_id.to_string()), None)
+            .await;
+        assert_eq!(resolved, known_id);
+    }
+
+    #[tokio::test]
+    async fn test_resolve_thread_with_none_external_thread_id_does_not_adopt() {
+        use crate::agent::session::Thread;
+
+        let manager = SessionManager::new();
+        let (session, default_tid) = manager.resolve_thread("user3", "chan3", None).await;
+
+        // Manually insert a thread with a known UUID (simulating a thread
+        // created by chat_new_thread_handler)
+        let known_id = Uuid::new_v4();
+        {
+            let mut sess = session.lock().await;
+            let thread = Thread::with_id(known_id, sess.id);
+            sess.threads.insert(known_id, thread);
+        }
+
+        // Resolve with external_thread_id=None but parsed_uuid=Some.
+        // This should NOT adopt the UUID — the old code prevented adoption
+        // when external_thread_id was None, and we preserve that invariant.
+        let (_, resolved) = manager
+            .resolve_thread_with_parsed_uuid("user3", "chan3", None, Some(known_id))
+            .await;
+
+        // Should return the existing default thread, not the injected UUID
+        assert_eq!(
+            resolved, default_tid,
+            "should return existing default thread when external_thread_id is None"
+        );
+        assert_ne!(
+            resolved, known_id,
+            "should NOT adopt UUID when external_thread_id is None"
+        );
+    }
 }

src/agent/submission.rs

@@ -92,6 +92,17 @@ impl SubmissionParser {
                 args: vec![],
             };
         }
+        if lower == "/reasoning" || lower.starts_with("/reasoning ") {
+            let args: Vec<String> = trimmed
+                .split_whitespace()
+                .skip(1)
+                .map(|s| s.to_string())
+                .collect();
+            return Submission::SystemCommand {
+                command: "reasoning".to_string(),
+                args,
+            };
+        }
         if lower == "/restart" {
             tracing::debug!("[SubmissionParser::parse] Recognized /restart command");
             return Submission::SystemCommand {

src/agent/thread_ops.rs

@@ -16,12 +16,12 @@ use crate::agent::dispatcher::{
 };
 use crate::agent::session::{MAX_PENDING_MESSAGES, PendingApproval, Session, ThreadState};
 use crate::agent::submission::SubmissionResult;
-use crate::channels::web::util::truncate_preview;
 use crate::channels::{IncomingMessage, StatusUpdate};
 use crate::context::JobContext;
 use crate::error::Error;
 use crate::llm::{ChatMessage, ToolCall};
 use crate::tools::redact_params;
+use ironclaw_common::truncate_preview;
 
 const FORGED_THREAD_ID_ERROR: &str = "Invalid or unauthorized thread ID.";
 
@@ -513,10 +513,10 @@ impl Agent {
                 };
 
                 thread.complete_turn(&response);
-                let (turn_number, tool_calls) = thread
+                let (turn_number, tool_calls, narrative) = thread
                     .turns
                     .last()
-                    .map(|t| (t.turn_number, t.tool_calls.clone()))
+                    .map(|t| (t.turn_number, t.tool_calls.clone(), t.narrative.clone()))
                     .unwrap_or_default();
                 let _ = self
                     .channels
@@ -534,6 +534,7 @@ impl Agent {
                     &message.user_id,
                     turn_number,
                     &tool_calls,
+                    narrative.as_deref(),
                 )
                 .await;
                 self.persist_assistant_response(
@@ -725,7 +726,9 @@ impl Agent {
     ///
     /// Stored between the user and assistant messages so that
     /// `build_turns_from_db_messages` can reconstruct the tool call history.
-    /// Content is a JSON array of tool call summaries.
+    /// Content is a JSON object: `{ "calls": [...], "narrative": "..." }`.
+    /// The `calls` array contains tool call summaries with optional `rationale`
+    /// and `tool_call_id` fields. Legacy rows may be plain JSON arrays.
     pub(super) async fn persist_tool_calls(
         &self,
         thread_id: Uuid,
@@ -733,6 +736,7 @@ impl Agent {
         user_id: &str,
         turn_number: usize,
         tool_calls: &[crate::agent::session::TurnToolCall],
+        narrative: Option<&str>,
     ) {
         if tool_calls.is_empty() {
             return;
@@ -767,11 +771,30 @@ impl Agent {
                 if let Some(ref error) = tc.error {
                     obj["error"] = serde_json::Value::String(truncate_preview(error, 200));
                 }
+                if let Some(ref rationale) = tc.rationale {
+                    obj["rationale"] = serde_json::Value::String(truncate_preview(rationale, 500));
+                }
+                if let Some(ref tool_call_id) = tc.tool_call_id {
+                    obj["tool_call_id"] =
+                        serde_json::Value::String(truncate_preview(tool_call_id, 128));
+                }
                 obj
             })
             .collect();
 
-        let content = match serde_json::to_string(&summaries) {
+        // Wrap in an object with optional narrative so it can be reconstructed.
+        // safety: no byte-index slicing here; comment describes JSON shape
+        let wrapper = if let Some(n) = narrative {
+            serde_json::json!({
+                "narrative": truncate_preview(n, 1000),
+                "calls": summaries,
+            })
+        } else {
+            serde_json::json!({
+                "calls": summaries,
+            })
+        };
+        let content = match serde_json::to_string(&wrapper) {
             Ok(c) => c,
             Err(e) => {
                 tracing::warn!("Failed to serialize tool calls: {}", e);
@@ -1104,9 +1127,12 @@ impl Agent {
                     && let Some(turn) = thread.last_turn_mut()
                 {
                     if is_tool_error {
-                        turn.record_tool_error(result_content.clone());
+                        turn.record_tool_error_for(&pending.tool_call_id, result_content.clone());
                     } else {
-                        turn.record_tool_result(serde_json::json!(result_content));
+                        turn.record_tool_result_for(
+                            &pending.tool_call_id,
+                            serde_json::json!(result_content),
+                        );
                     }
                 }
             }
@@ -1358,9 +1384,12 @@ impl Agent {
                         && let Some(turn) = thread.last_turn_mut()
                     {
                         if is_deferred_error {
-                            turn.record_tool_error(deferred_content.clone());
+                            turn.record_tool_error_for(&tc.id, deferred_content.clone());
                         } else {
-                            turn.record_tool_result(serde_json::json!(deferred_content));
+                            turn.record_tool_result_for(
+                                &tc.id,
+                                serde_json::json!(deferred_content),
+                            );
                         }
                     }
                 }
@@ -1459,10 +1488,10 @@ impl Agent {
                     let (response, suggestions) =
                         crate::agent::dispatcher::extract_suggestions(&response);
                     thread.complete_turn(&response);
-                    let (turn_number, tool_calls) = thread
+                    let (turn_number, tool_calls, narrative) = thread
                         .turns
                         .last()
-                        .map(|t| (t.turn_number, t.tool_calls.clone()))
+                        .map(|t| (t.turn_number, t.tool_calls.clone(), t.narrative.clone()))
                         .unwrap_or_default();
                     // User message already persisted at turn start; save tool calls then assistant response
                     self.persist_tool_calls(
@@ -1471,6 +1500,7 @@ impl Agent {
                         &message.user_id,
                         turn_number,
                         &tool_calls,
+                        narrative.as_deref(),
                     )
                     .await;
                     self.persist_assistant_response(
@@ -1646,7 +1676,7 @@ impl Agent {
         };
 
         match ext_mgr
-            .configure_token(&pending.extension_name, token)
+            .configure_token(&pending.extension_name, token, &message.user_id)
             .await
         {
             Ok(result) if result.activated => {
@@ -1816,7 +1846,20 @@ fn rebuild_chat_messages_from_db(
             "assistant" => result.push(ChatMessage::assistant(&msg.content)),
             "tool_calls" => {
                 // Try to parse the enriched JSON and rebuild tool messages.
-                if let Ok(calls) = serde_json::from_str::<Vec<serde_json::Value>>(&msg.content) {
+                // Supports two formats:
+                // - Old: plain JSON array of tool call summaries
+                // - New: wrapped object { "calls": [...], "narrative": "..." }
+                let calls: Vec<serde_json::Value> =
+                    match serde_json::from_str::<serde_json::Value>(&msg.content) {
+                        Ok(serde_json::Value::Array(arr)) => arr,
+                        Ok(serde_json::Value::Object(obj)) => obj
+                            .get("calls")
+                            .and_then(|v| v.as_array())
+                            .cloned()
+                            .unwrap_or_default(),
+                        _ => Vec::new(),
+                    };
+                {
                     if calls.is_empty() {
                         continue;
                     }
@@ -1839,6 +1882,10 @@ fn rebuild_chat_messages_from_db(
                                     .get("parameters")
                                     .cloned()
                                     .unwrap_or(serde_json::json!({})),
+                                reasoning: c
+                                    .get("rationale")
+                                    .and_then(|v| v.as_str())
+                                    .map(String::from),
                             })
                             .collect();
 

src/app.rs

@@ -312,13 +312,7 @@ impl AppBuilder {
             .create_provider(&self.config.llm.nearai.base_url, self.session.clone());
 
         // Register memory tools if database is available
-        let workspace_user_id = self
+        let workspace_user_id = self.config.owner_id.as_str();
-            .config
-            .channels
-            .gateway
-            .as_ref()
-            .map(|gw| gw.user_id.as_str())
-            .unwrap_or("default");
         let workspace = if let Some(ref db) = self.db {
             let emb_cache_config = EmbeddingCacheConfig {
                 max_entries: self.config.embeddings.cache_size,
@@ -327,7 +321,7 @@ impl AppBuilder {
                 .with_search_config(&self.config.search);
 
             if let Some(ref emb) = embeddings {
-                ws = ws.with_embeddings_cached(emb.clone(), emb_cache_config);
+                ws = ws.with_embeddings_cached(emb.clone(), emb_cache_config.clone());
             }
 
             // Wire workspace-level settings (read scopes, memory layers)
@@ -341,7 +335,35 @@ impl AppBuilder {
             }
             ws = ws.with_memory_layers(self.config.workspace.memory_layers.clone());
             let ws = Arc::new(ws);
-            tools.register_memory_tools(Arc::clone(&ws));
+
+            // Detect multi-tenant mode: when GATEWAY_USER_TOKENS is configured,
+            // each authenticated user needs their own workspace scope. Use
+            // WorkspacePool (which implements WorkspaceResolver) to create
+            // per-user workspaces on demand instead of sharing the startup
+            // workspace across all users.
+            let is_multi_tenant = self
+                .config
+                .channels
+                .gateway
+                .as_ref()
+                .is_some_and(|gw| gw.user_tokens.is_some());
+
+            if is_multi_tenant {
+                let pool = Arc::new(crate::channels::web::server::WorkspacePool::new(
+                    Arc::clone(db),
+                    embeddings.clone(),
+                    emb_cache_config,
+                    self.config.search.clone(),
+                    self.config.workspace.clone(),
+                ));
+                tools.register_memory_tools_with_resolver(pool);
+                tracing::info!(
+                    "Memory tools configured with per-user workspace resolver (multi-tenant mode)"
+                );
+            } else {
+                tools.register_memory_tools(Arc::clone(&ws));
+            }
+
             Some(ws)
         } else {
             None

src/channels/channel.rs

@@ -265,6 +265,15 @@ impl OutgoingResponse {
     }
 }
 
+/// A single tool decision within a reasoning update.
+#[derive(Debug, Clone)]
+pub struct ToolDecision {
+    /// Tool name.
+    pub tool_name: String,
+    /// Agent's reasoning for choosing this tool.
+    pub rationale: String,
+}
+
 /// Status update types for showing agent activity.
 #[derive(Debug, Clone)]
 pub enum StatusUpdate {
@@ -333,6 +342,13 @@ pub enum StatusUpdate {
     },
     /// Suggested follow-up messages for the user.
     Suggestions { suggestions: Vec<String> },
+    /// Agent reasoning update (why it chose specific tools).
+    ReasoningUpdate {
+        /// Human-readable summary of the agent's decision.
+        narrative: String,
+        /// Per-tool decisions.
+        decisions: Vec<ToolDecision>,
+    },
     /// Per-turn token usage and cost summary (shown as subtle metadata).
     TurnCost {
         input_tokens: u64,

src/channels/mod.rs

@@ -39,7 +39,7 @@ mod webhook_server;
 
 pub use channel::{
     AttachmentKind, Channel, ChannelSecretUpdater, IncomingAttachment, IncomingMessage,
-    MessageStream, OutgoingResponse, StatusUpdate, routing_target_from_metadata,
+    MessageStream, OutgoingResponse, StatusUpdate, ToolDecision, routing_target_from_metadata,
 };
 pub use http::{HttpChannel, HttpChannelState};
 pub use manager::ChannelManager;

src/channels/repl.rs

@@ -75,6 +75,7 @@ const SLASH_COMMANDS: &[&str] = &[
     "/suggest",
     "/thread",
     "/resume",
+    "/reasoning",
 ];
 
 /// Rustyline helper for slash-command tab completion.
@@ -430,6 +431,18 @@ impl ReplChannel {
             let _ = execute!(stderr, terminal::Clear(terminal::ClearType::FromCursorDown));
         }
     }
+
+    async fn finish_single_message_turn(&self) {
+        if self.single_message.is_none() {
+            return;
+        }
+
+        let tx = self.msg_tx.lock().ok().and_then(|mut guard| guard.take());
+        if let Some(tx) = tx {
+            let msg = IncomingMessage::new("repl", &self.user_id, "/quit");
+            let _ = tx.send(msg).await;
+        }
+    }
 }
 
 impl Default for ReplChannel {
@@ -479,7 +492,9 @@ impl Channel for ReplChannel {
 
     async fn start(&self) -> Result<MessageStream, ChannelError> {
         let (tx, rx) = mpsc::channel(32);
-        // Store tx so send_status can inject approval responses directly
+        // Approval prompts inject responses back through this sender.
+        // In single-message mode we keep it until the turn finishes, then
+        // drop it after enqueuing /quit so the receiver stream can close.
         if let Ok(mut guard) = self.msg_tx.lock() {
             *guard = Some(tx.clone());
         }
@@ -495,11 +510,10 @@ impl Channel for ReplChannel {
 
             // Single message mode: send it and return
             if let Some(msg) = single_message {
-                let incoming = IncomingMessage::new("repl", &user_id, &msg).with_timezone(&sys_tz);
+                let incoming = IncomingMessage::new("repl", &user_id, &msg)
+                    .with_metadata(serde_json::json!({ "single_message_mode": true }))
+                    .with_timezone(&sys_tz);
                 let _ = tx.blocking_send(incoming);
-                // Ensure the agent exits after handling exactly one turn in -m mode,
-                // even when other channels (gateway/http) are enabled.
-                let _ = tx.blocking_send(IncomingMessage::new("repl", &user_id, "/quit"));
                 return;
             }
 
@@ -662,6 +676,7 @@ impl Channel for ReplChannel {
             println!();
             println!();
             self.stdin_locked.store(false, Ordering::Relaxed);
+            self.finish_single_message_turn().await;
             return Ok(());
         }
 
@@ -680,6 +695,7 @@ impl Channel for ReplChannel {
         println!();
         // Unlock stdin so readline can resume
         self.stdin_locked.store(false, Ordering::Relaxed);
+        self.finish_single_message_turn().await;
         Ok(())
     }
 
@@ -779,6 +795,7 @@ impl Channel for ReplChannel {
                 let msg_tx = Arc::clone(&self.msg_tx);
                 let user_id = self.user_id.clone();
                 let lock_flag = Arc::clone(&self.stdin_locked);
+                let single_message_mode = self.single_message.is_some();
                 tokio::task::spawn_blocking(move || {
                     let action = run_approval_selector(allow_always).unwrap_or("n");
                     // Unlock stdin so readline can resume after approval
@@ -787,7 +804,12 @@ impl Channel for ReplChannel {
                         return;
                     };
                     if let Some(tx) = guard.as_ref() {
-                        let msg = IncomingMessage::new("repl", &user_id, action);
+                        let msg = if single_message_mode {
+                            IncomingMessage::new("repl", &user_id, action)
+                                .with_metadata(serde_json::json!({ "single_message_mode": true }))
+                        } else {
+                            IncomingMessage::new("repl", &user_id, action)
+                        };
                         let _ = tx.blocking_send(msg);
                     }
                 });
@@ -841,6 +863,19 @@ impl Channel for ReplChannel {
             StatusUpdate::Suggestions { .. } => {
                 // Suggestions are only rendered by the web gateway
             }
+            StatusUpdate::ReasoningUpdate {
+                narrative,
+                decisions,
+            } => {
+                if !narrative.is_empty() {
+                    let display = truncate_for_preview(&narrative, CLI_STATUS_MAX);
+                    eprintln!("  \x1b[94m\u{25B6} {display}\x1b[0m");
+                }
+                for d in &decisions {
+                    let display = truncate_for_preview(&d.rationale, CLI_STATUS_MAX);
+                    eprintln!("    \x1b[90m\u{2192} {}: {display}\x1b[0m", d.tool_name);
+                }
+            }
             StatusUpdate::TurnCost { .. } => {
                 // Cost display is handled by the TUI channel
             }
@@ -875,6 +910,7 @@ impl Channel for ReplChannel {
 #[cfg(test)]
 mod tests {
     use futures::StreamExt;
+    use tokio::time::{Duration, timeout};
 
     use super::*;
 
@@ -883,16 +919,36 @@ mod tests {
         let repl = ReplChannel::with_message("hi".to_string());
         let mut stream = repl.start().await.expect("repl start should succeed");
 
-        let first = stream.next().await.expect("first message missing");
+        let first = timeout(Duration::from_secs(1), stream.next())
+            .await
+            .expect("timed out waiting for first message")
+            .expect("first message missing");
         assert_eq!(first.channel, "repl");
         assert_eq!(first.content, "hi");
 
-        let second = stream.next().await.expect("quit message missing");
+        assert!(
+            timeout(Duration::from_millis(100), stream.next())
+                .await
+                .is_err(),
+            "single-message mode should wait for the turn to finish before quitting"
+        );
+
+        repl.respond(&first, OutgoingResponse::text("done"))
+            .await
+            .expect("respond should succeed");
+
+        let second = timeout(Duration::from_secs(1), stream.next())
+            .await
+            .expect("timed out waiting for quit message")
+            .expect("quit message missing");
         assert_eq!(second.channel, "repl");
         assert_eq!(second.content, "/quit");
 
         assert!(
-            stream.next().await.is_none(),
+            timeout(Duration::from_secs(1), stream.next())
+                .await
+                .expect("timed out waiting for stream to close")
+                .is_none(),
             "stream should end after /quit"
         );
     }

src/channels/wasm/router.rs

@@ -333,6 +333,9 @@ async fn webhook_handler(
 
     let channel_name = channel.channel_name();
 
+    // Track whether any authentication was performed and passed.
+    let mut did_authenticate = false;
+
     // Check if secret is required
     if state.router.requires_secret(channel_name).await {
         // Get the secret header name for this channel (from capabilities or default)
@@ -382,6 +385,7 @@ async fn webhook_handler(
                     );
                 }
                 tracing::debug!(channel = %channel_name, "Webhook secret validated");
+                did_authenticate = true;
             }
             None => {
                 tracing::warn!(
@@ -433,6 +437,7 @@ async fn webhook_handler(
                     );
                 }
                 tracing::debug!(channel = %channel_name, "Ed25519 signature verified");
+                did_authenticate = true;
             }
             _ => {
                 tracing::warn!(
@@ -484,6 +489,7 @@ async fn webhook_handler(
                     );
                 }
                 tracing::debug!(channel = %channel_name, "HMAC-SHA256 signature verified");
+                did_authenticate = true;
             }
             _ => {
                 tracing::warn!(
@@ -510,8 +516,9 @@ async fn webhook_handler(
         })
         .collect();
 
-    // Call the WASM channel
+    // Call the WASM channel. `did_authenticate` was set above by whichever
-    let secret_validated = state.router.requires_secret(channel_name).await;
+    // auth guard (secret / Ed25519 / HMAC) successfully validated the request.
+    let secret_validated = did_authenticate;
 
     tracing::info!(
         channel = %channel_name,

src/channels/wasm/wrapper.rs

@@ -3061,6 +3061,20 @@ fn status_to_wit(
         },
         // Suggestions and turn cost are web-gateway-only; skip for WASM channels
         StatusUpdate::Suggestions { .. } | StatusUpdate::TurnCost { .. } => return None,
+        StatusUpdate::ReasoningUpdate {
+            narrative,
+            decisions,
+        } => {
+            let mut msg = narrative.clone();
+            for d in decisions {
+                msg.push_str(&format!("\n  → {}: {}", d.tool_name, d.rationale));
+            }
+            wit_channel::StatusUpdate {
+                status: wit_channel::StatusType::Status,
+                message: msg,
+                metadata_json,
+            }
+        }
     })
 }
 

src/channels/web/auth.rs

@@ -1,17 +1,133 @@
 //! Bearer token authentication middleware for the web gateway.
+//!
+//! Supports multi-user mode: each token maps to a `UserIdentity` that carries
+//! the user_id. The identity is inserted into request extensions so downstream
+//! handlers can extract it via `AuthenticatedUser`.
+
+use std::collections::HashMap;
 
 use axum::{
-    extract::{Request, State},
+    extract::{FromRequestParts, Request, State},
-    http::{HeaderMap, Method, StatusCode},
+    http::{HeaderMap, Method, StatusCode, request::Parts},
     middleware::Next,
     response::{IntoResponse, Response},
 };
+use sha2::{Digest, Sha256};
 use subtle::ConstantTimeEq;
 
-/// Shared auth state injected via axum middleware state.
+/// Identity resolved from a bearer token.
+#[derive(Debug, Clone)]
+pub struct UserIdentity {
+    pub user_id: String,
+    /// Additional user scopes this identity can read from.
+    pub workspace_read_scopes: Vec<String>,
+}
+
+/// Hash a token with SHA-256 for constant-size, timing-safe storage.
+fn hash_token(token: &str) -> [u8; 32] {
+    let mut hasher = Sha256::new();
+    hasher.update(token.as_bytes());
+    hasher.finalize().into()
+}
+
+/// Multi-user auth state: maps token hashes to user identities.
+///
+/// Tokens are SHA-256 hashed on construction so they are never stored in
+/// plaintext. Authentication compares fixed-size (32-byte) digests using
+/// constant-time comparison, eliminating both length-oracle timing leaks
+/// and accidental token exposure in memory dumps.
+///
+/// In single-user mode (the default), contains exactly one entry.
 #[derive(Clone)]
-pub struct AuthState {
+pub struct MultiAuthState {
-    pub token: String,
+    /// Maps SHA-256(token) → identity. Tokens are never stored in cleartext.
+    hashed_tokens: Vec<([u8; 32], UserIdentity)>,
+    /// Original first token kept only for single-user startup printing.
+    /// Not used for authentication.
+    display_token: Option<String>,
+}
+
+impl MultiAuthState {
+    /// Create a single-user auth state (backwards compatible).
+    pub fn single(token: String, user_id: String) -> Self {
+        let hash = hash_token(&token);
+        Self {
+            hashed_tokens: vec![(
+                hash,
+                UserIdentity {
+                    user_id,
+                    workspace_read_scopes: Vec::new(),
+                },
+            )],
+            display_token: Some(token),
+        }
+    }
+
+    /// Create a multi-user auth state from a map of tokens to identities.
+    pub fn multi(tokens: HashMap<String, UserIdentity>) -> Self {
+        let hashed_tokens: Vec<([u8; 32], UserIdentity)> = tokens
+            .into_iter()
+            .map(|(tok, identity)| (hash_token(&tok), identity))
+            .collect();
+        Self {
+            hashed_tokens,
+            display_token: None,
+        }
+    }
+
+    /// Authenticate a token, returning the associated identity if valid.
+    ///
+    /// Uses SHA-256 hashing + constant-time comparison (`subtle::ConstantTimeEq`)
+    /// to prevent timing side-channels. Both the candidate and stored tokens are
+    /// hashed to 32-byte digests, eliminating length-oracle leaks. Iterates all
+    /// entries regardless of match to avoid early-exit timing differences.
+    /// O(n) in the number of configured users — negligible for typical
+    /// deployments (< 10 users).
+    pub fn authenticate(&self, candidate: &str) -> Option<&UserIdentity> {
+        let candidate_hash = hash_token(candidate);
+        let mut matched: Option<&UserIdentity> = None;
+        for (stored_hash, identity) in &self.hashed_tokens {
+            if bool::from(candidate_hash.ct_eq(stored_hash)) {
+                matched = Some(identity);
+            }
+        }
+        matched
+    }
+
+    /// Get the first token for backwards-compatible printing at startup.
+    ///
+    /// Only available in single-user mode; returns `None` in multi-user mode
+    /// to avoid exposing tokens.
+    pub fn first_token(&self) -> Option<&str> {
+        self.display_token.as_deref()
+    }
+
+    /// Get the first user identity (for single-user fallback).
+    pub fn first_identity(&self) -> Option<&UserIdentity> {
+        self.hashed_tokens.first().map(|(_, id)| id)
+    }
+}
+
+/// Axum extractor that provides the authenticated user identity.
+///
+/// Only available on routes behind `auth_middleware`. Extracts the
+/// `UserIdentity` that the middleware inserted into request extensions.
+pub struct AuthenticatedUser(pub UserIdentity);
+
+impl<S> FromRequestParts<S> for AuthenticatedUser
+where
+    S: Send + Sync,
+{
+    type Rejection = (StatusCode, &'static str);
+
+    async fn from_request_parts(parts: &mut Parts, _state: &S) -> Result<Self, Self::Rejection> {
+        parts
+            .extensions
+            .get::<UserIdentity>()
+            .cloned()
+            .map(AuthenticatedUser)
+            .ok_or((StatusCode::UNAUTHORIZED, "Not authenticated"))
+    }
 }
 
 /// Whether query-string token auth is allowed for this request.
@@ -51,29 +167,34 @@ fn query_token(request: &Request) -> Option<String> {
 /// Auth middleware that validates bearer token from header or query param.
 ///
 /// SSE connections can't set headers from `EventSource`, so we also accept
-/// `?token=xxx` as a query parameter, but only on SSE endpoints.
+/// `?token=xxx` as a query parameter, but only on SSE/WS endpoints.
+///
+/// On successful authentication, inserts the matching `UserIdentity` into
+/// request extensions for downstream extraction via `AuthenticatedUser`.
 pub async fn auth_middleware(
-    State(auth): State<AuthState>,
+    State(auth): State<MultiAuthState>,
     headers: HeaderMap,
-    request: Request,
+    mut request: Request,
     next: Next,
 ) -> Response {
-    // Try Authorization header first (constant-time comparison).
+    // Try Authorization header first.
     // RFC 6750 Section 2.1: auth-scheme comparison is case-insensitive.
     if let Some(auth_header) = headers.get("authorization")
         && let Ok(value) = auth_header.to_str()
         && value.len() > 7
         && value[..7].eq_ignore_ascii_case("Bearer ")
-        && bool::from(value.as_bytes()[7..].ct_eq(auth.token.as_bytes()))
+        && let Some(identity) = auth.authenticate(&value[7..])
     {
+        request.extensions_mut().insert(identity.clone());
         return next.run(request).await;
     }
 
-    // Fall back to query parameter, but only for SSE endpoints (constant-time comparison).
+    // Fall back to query parameter, but only for SSE/WS endpoints.
     if allows_query_token_auth(&request)
         && let Some(token) = query_token(&request)
-        && bool::from(token.as_bytes().ct_eq(auth.token.as_bytes()))
+        && let Some(identity) = auth.authenticate(&token)
     {
+        request.extensions_mut().insert(identity.clone());
         return next.run(request).await;
     }
 
@@ -83,15 +204,61 @@ pub async fn auth_middleware(
 #[cfg(test)]
 mod tests {
     use super::*;
-    use crate::testing::credentials::{TEST_AUTH_SECRET_TOKEN, TEST_BEARER_TOKEN};
+    use crate::testing::credentials::TEST_AUTH_SECRET_TOKEN;
 
     #[test]
-    fn test_auth_state_clone() {
+    fn test_multi_auth_state_single() {
-        let state = AuthState {
+        let state = MultiAuthState::single("tok-123".to_string(), "alice".to_string());
-            token: TEST_BEARER_TOKEN.to_string(),
+        let identity = state.authenticate("tok-123");
-        };
+        assert!(identity.is_some());
-        let cloned = state.clone();
+        assert_eq!(identity.unwrap().user_id, "alice");
-        assert_eq!(cloned.token, TEST_BEARER_TOKEN);
+    }
+
+    #[test]
+    fn test_multi_auth_state_reject_wrong_token() {
+        let state = MultiAuthState::single("tok-123".to_string(), "alice".to_string());
+        assert!(state.authenticate("wrong-token").is_none());
+    }
+
+    #[test]
+    fn test_multi_auth_state_multi_users() {
+        let mut tokens = HashMap::new();
+        tokens.insert(
+            "tok-alice".to_string(),
+            UserIdentity {
+                user_id: "alice".to_string(),
+                workspace_read_scopes: Vec::new(),
+            },
+        );
+        tokens.insert(
+            "tok-bob".to_string(),
+            UserIdentity {
+                user_id: "bob".to_string(),
+                workspace_read_scopes: Vec::new(),
+            },
+        );
+        let state = MultiAuthState::multi(tokens);
+
+        let alice = state.authenticate("tok-alice").unwrap();
+        assert_eq!(alice.user_id, "alice");
+
+        let bob = state.authenticate("tok-bob").unwrap();
+        assert_eq!(bob.user_id, "bob");
+
+        assert!(state.authenticate("tok-charlie").is_none());
+    }
+
+    #[test]
+    fn test_multi_auth_state_first_token() {
+        let state = MultiAuthState::single("my-token".to_string(), "user1".to_string());
+        assert_eq!(state.first_token(), Some("my-token"));
+    }
+
+    #[test]
+    fn test_multi_auth_state_first_identity() {
+        let state = MultiAuthState::single("my-token".to_string(), "user1".to_string());
+        let identity = state.first_identity().unwrap();
+        assert_eq!(identity.user_id, "user1");
     }
 
     use axum::Router;
@@ -107,9 +274,7 @@ mod tests {
     /// Router with streaming endpoints (query auth allowed) and regular
     /// endpoints (query auth rejected).
     fn test_app(token: &str) -> Router {
-        let state = AuthState {
+        let state = MultiAuthState::single(token.to_string(), "test-user".to_string());
-            token: token.to_string(),
-        };
         Router::new()
             .route("/api/chat/events", get(dummy_handler))
             .route("/api/logs/events", get(dummy_handler))
@@ -306,4 +471,200 @@ mod tests {
         let resp = app.oneshot(req).await.unwrap();
         assert_eq!(resp.status(), StatusCode::UNAUTHORIZED);
     }
+
+    // --- Multi-tenant auth integration tests ---
+
+    /// Handler that extracts `AuthenticatedUser` and returns the resolved user_id.
+    async fn identity_handler(AuthenticatedUser(identity): AuthenticatedUser) -> String {
+        identity.user_id
+    }
+
+    /// Handler that extracts `AuthenticatedUser` and returns workspace_read_scopes as JSON.
+    async fn scopes_handler(AuthenticatedUser(identity): AuthenticatedUser) -> String {
+        serde_json::to_string(&identity.workspace_read_scopes).unwrap()
+    }
+
+    /// Build a multi-user router where each token maps to a distinct identity.
+    fn multi_user_app(tokens: HashMap<String, UserIdentity>) -> Router {
+        let state = MultiAuthState::multi(tokens);
+        Router::new()
+            .route("/api/chat/events", get(identity_handler))
+            .route("/api/chat/send", post(identity_handler))
+            .route("/api/scopes", get(scopes_handler))
+            .layer(middleware::from_fn_with_state(state, auth_middleware))
+    }
+
+    fn two_user_tokens() -> HashMap<String, UserIdentity> {
+        let mut tokens = HashMap::new();
+        tokens.insert(
+            "tok-alice".to_string(),
+            UserIdentity {
+                user_id: "alice".to_string(),
+                workspace_read_scopes: vec!["shared".to_string()],
+            },
+        );
+        tokens.insert(
+            "tok-bob".to_string(),
+            UserIdentity {
+                user_id: "bob".to_string(),
+                workspace_read_scopes: vec!["shared".to_string(), "alice".to_string()],
+            },
+        );
+        tokens
+    }
+
+    #[tokio::test]
+    async fn test_multi_user_alice_token_resolves_to_alice() {
+        let app = multi_user_app(two_user_tokens());
+        let req = Request::builder()
+            .uri("/api/chat/events")
+            .header("Authorization", "Bearer tok-alice")
+            .body(Body::empty())
+            .unwrap();
+        let resp = app.oneshot(req).await.unwrap();
+        assert_eq!(resp.status(), StatusCode::OK);
+        let body = axum::body::to_bytes(resp.into_body(), 1024).await.unwrap();
+        assert_eq!(body, "alice");
+    }
+
+    #[tokio::test]
+    async fn test_multi_user_bob_token_resolves_to_bob() {
+        let app = multi_user_app(two_user_tokens());
+        let req = Request::builder()
+            .uri("/api/chat/events")
+            .header("Authorization", "Bearer tok-bob")
+            .body(Body::empty())
+            .unwrap();
+        let resp = app.oneshot(req).await.unwrap();
+        assert_eq!(resp.status(), StatusCode::OK);
+        let body = axum::body::to_bytes(resp.into_body(), 1024).await.unwrap();
+        assert_eq!(body, "bob");
+    }
+
+    #[tokio::test]
+    async fn test_multi_user_sequential_tokens_resolve_independently() {
+        // Send both alice and bob tokens sequentially and verify each gets
+        // the correct identity — guards against token map corruption.
+        let tokens = two_user_tokens();
+
+        let app1 = multi_user_app(tokens.clone());
+        let req = Request::builder()
+            .uri("/api/chat/events")
+            .header("Authorization", "Bearer tok-alice")
+            .body(Body::empty())
+            .unwrap();
+        let resp = app1.oneshot(req).await.unwrap();
+        let body = axum::body::to_bytes(resp.into_body(), 1024).await.unwrap();
+        assert_eq!(body, "alice");
+
+        let app2 = multi_user_app(tokens);
+        let req = Request::builder()
+            .uri("/api/chat/events")
+            .header("Authorization", "Bearer tok-bob")
+            .body(Body::empty())
+            .unwrap();
+        let resp = app2.oneshot(req).await.unwrap();
+        let body = axum::body::to_bytes(resp.into_body(), 1024).await.unwrap();
+        assert_eq!(body, "bob");
+    }
+
+    #[tokio::test]
+    async fn test_multi_user_unknown_token_rejected() {
+        let app = multi_user_app(two_user_tokens());
+        let req = Request::builder()
+            .uri("/api/chat/events")
+            .header("Authorization", "Bearer tok-charlie")
+            .body(Body::empty())
+            .unwrap();
+        let resp = app.oneshot(req).await.unwrap();
+        assert_eq!(resp.status(), StatusCode::UNAUTHORIZED);
+    }
+
+    #[tokio::test]
+    async fn test_multi_user_workspace_read_scopes_propagated() {
+        let app = multi_user_app(two_user_tokens());
+
+        // Alice has ["shared"]
+        let req = Request::builder()
+            .uri("/api/scopes")
+            .header("Authorization", "Bearer tok-alice")
+            .body(Body::empty())
+            .unwrap();
+        let resp = app.oneshot(req).await.unwrap();
+        let body = axum::body::to_bytes(resp.into_body(), 1024).await.unwrap();
+        let scopes: Vec<String> = serde_json::from_slice(&body).unwrap();
+        assert_eq!(scopes, vec!["shared"]);
+    }
+
+    #[tokio::test]
+    async fn test_multi_user_bob_has_two_scopes() {
+        let app = multi_user_app(two_user_tokens());
+
+        // Bob has ["shared", "alice"]
+        let req = Request::builder()
+            .uri("/api/scopes")
+            .header("Authorization", "Bearer tok-bob")
+            .body(Body::empty())
+            .unwrap();
+        let resp = app.oneshot(req).await.unwrap();
+        let body = axum::body::to_bytes(resp.into_body(), 1024).await.unwrap();
+        let scopes: Vec<String> = serde_json::from_slice(&body).unwrap();
+        assert_eq!(scopes, vec!["shared", "alice"]);
+    }
+
+    #[tokio::test]
+    async fn test_multi_user_query_param_resolves_correct_identity() {
+        let app = multi_user_app(two_user_tokens());
+        let req = Request::builder()
+            .uri("/api/chat/events?token=tok-bob")
+            .body(Body::empty())
+            .unwrap();
+        let resp = app.oneshot(req).await.unwrap();
+        assert_eq!(resp.status(), StatusCode::OK);
+        let body = axum::body::to_bytes(resp.into_body(), 1024).await.unwrap();
+        assert_eq!(body, "bob");
+    }
+
+    #[tokio::test]
+    async fn test_multi_user_post_with_bearer_resolves_identity() {
+        let app = multi_user_app(two_user_tokens());
+        let req = Request::builder()
+            .method(Method::POST)
+            .uri("/api/chat/send")
+            .header("Authorization", "Bearer tok-alice")
+            .body(Body::empty())
+            .unwrap();
+        let resp = app.oneshot(req).await.unwrap();
+        assert_eq!(resp.status(), StatusCode::OK);
+        let body = axum::body::to_bytes(resp.into_body(), 1024).await.unwrap();
+        assert_eq!(body, "alice");
+    }
+
+    #[tokio::test]
+    async fn test_multi_user_empty_scopes_for_single_user() {
+        // Single-user mode creates identity with empty workspace_read_scopes.
+        let state = MultiAuthState::single("tok-only".to_string(), "solo".to_string());
+        let app = Router::new()
+            .route("/api/scopes", get(scopes_handler))
+            .layer(middleware::from_fn_with_state(state, auth_middleware));
+        let req = Request::builder()
+            .uri("/api/scopes")
+            .header("Authorization", "Bearer tok-only")
+            .body(Body::empty())
+            .unwrap();
+        let resp = app.oneshot(req).await.unwrap();
+        let body = axum::body::to_bytes(resp.into_body(), 1024).await.unwrap();
+        let scopes: Vec<String> = serde_json::from_slice(&body).unwrap();
+        assert!(scopes.is_empty());
+    }
+
+    #[tokio::test]
+    async fn test_prefix_and_extension_tokens_rejected() {
+        // Verifies that prefix/suffix variants of valid tokens are rejected.
+        // Note: the constant-time property is enforced structurally by use of
+        // subtle::ConstantTimeEq and cannot be verified via outcome testing.
+        let state = MultiAuthState::single("long-secret-token".to_string(), "user".to_string());
+        assert!(state.authenticate("long-secret").is_none());
+        assert!(state.authenticate("long-secret-token-extra").is_none());
+    }
 }

src/channels/web/handlers/chat.rs

@@ -12,22 +12,24 @@ use serde::Deserialize;
 use uuid::Uuid;
 
 use crate::channels::IncomingMessage;
+use crate::channels::web::auth::AuthenticatedUser;
 use crate::channels::web::server::GatewayState;
 use crate::channels::web::types::*;
 use crate::channels::web::util::{build_turns_from_db_messages, truncate_preview};
 
 pub async fn chat_send_handler(
     State(state): State<Arc<GatewayState>>,
+    AuthenticatedUser(identity): AuthenticatedUser,
     Json(req): Json<SendMessageRequest>,
 ) -> Result<(StatusCode, Json<SendMessageResponse>), (StatusCode, String)> {
-    if !state.chat_rate_limiter.check() {
+    if !state.chat_rate_limiter.check(&identity.user_id) {
         return Err((
             StatusCode::TOO_MANY_REQUESTS,
             "Rate limit exceeded. Try again shortly.".to_string(),
         ));
     }
 
-    let mut msg = IncomingMessage::new("gateway", &state.user_id, &req.content);
+    let mut msg = IncomingMessage::new("gateway", &identity.user_id, &req.content);
 
     if let Some(ref thread_id) = req.thread_id {
         msg = msg.with_thread(thread_id);
@@ -74,6 +76,7 @@ pub async fn chat_send_handler(
 
 pub async fn chat_approval_handler(
     State(state): State<Arc<GatewayState>>,
+    AuthenticatedUser(identity): AuthenticatedUser,
     Json(req): Json<ApprovalRequest>,
 ) -> Result<(StatusCode, Json<SendMessageResponse>), (StatusCode, String)> {
     let (approved, always) = match req.action.as_str() {
@@ -109,7 +112,7 @@ pub async fn chat_approval_handler(
         )
     })?;
 
-    let mut msg = IncomingMessage::new("gateway", &state.user_id, content);
+    let mut msg = IncomingMessage::new("gateway", &identity.user_id, content);
 
     if let Some(ref thread_id) = req.thread_id {
         msg = msg.with_thread(thread_id);
@@ -150,6 +153,7 @@ pub async fn chat_approval_handler(
 /// The token never touches the LLM, chat history, or SSE stream.
 pub async fn chat_auth_token_handler(
     State(state): State<Arc<GatewayState>>,
+    AuthenticatedUser(user): AuthenticatedUser,
     Json(req): Json<AuthTokenRequest>,
 ) -> Result<Json<ActionResponse>, (StatusCode, String)> {
     let ext_mgr = state.extension_manager.as_ref().ok_or((
@@ -158,7 +162,7 @@ pub async fn chat_auth_token_handler(
     ))?;
 
     match ext_mgr
-        .configure_token(&req.extension_name, &req.token)
+        .configure_token(&req.extension_name, &req.token, &user.user_id)
         .await
     {
         Ok(result) => {
@@ -169,20 +173,26 @@ pub async fn chat_auth_token_handler(
             resp.instructions = result.verification.as_ref().map(|v| v.instructions.clone());
 
             if result.verification.is_some() {
-                state.sse.broadcast(SseEvent::AuthRequired {
+                state.sse.broadcast_for_user(
-                    extension_name: req.extension_name.clone(),
+                    &user.user_id,
-                    instructions: Some(result.message),
+                    AppEvent::AuthRequired {
-                    auth_url: None,
+                        extension_name: req.extension_name.clone(),
-                    setup_url: None,
+                        instructions: Some(result.message),
-                });
+                        auth_url: None,
+                        setup_url: None,
+                    },
+                );
             } else {
-                clear_auth_mode(&state).await;
+                clear_auth_mode(&state, &user.user_id).await;
 
-                state.sse.broadcast(SseEvent::AuthCompleted {
+                state.sse.broadcast_for_user(
-                    extension_name: req.extension_name.clone(),
+                    &user.user_id,
-                    success: true,
+                    AppEvent::AuthCompleted {
-                    message: result.message,
+                        extension_name: req.extension_name.clone(),
-                });
+                        success: true,
+                        message: result.message,
+                    },
+                );
             }
 
             Ok(Json(resp))
@@ -190,12 +200,15 @@ pub async fn chat_auth_token_handler(
         Err(e) => {
             let msg = e.to_string();
             if matches!(e, crate::extensions::ExtensionError::ValidationFailed(_)) {
-                state.sse.broadcast(SseEvent::AuthRequired {
+                state.sse.broadcast_for_user(
-                    extension_name: req.extension_name.clone(),
+                    &user.user_id,
-                    instructions: Some(msg.clone()),
+                    AppEvent::AuthRequired {
-                    auth_url: None,
+                        extension_name: req.extension_name.clone(),
-                    setup_url: None,
+                        instructions: Some(msg.clone()),
-                });
+                        auth_url: None,
+                        setup_url: None,
+                    },
+                );
             }
             Ok(Json(ActionResponse::fail(msg)))
         }
@@ -205,16 +218,17 @@ pub async fn chat_auth_token_handler(
 /// Cancel an in-progress auth flow.
 pub async fn chat_auth_cancel_handler(
     State(state): State<Arc<GatewayState>>,
+    AuthenticatedUser(identity): AuthenticatedUser,
     Json(_req): Json<AuthCancelRequest>,
 ) -> Result<Json<ActionResponse>, (StatusCode, String)> {
-    clear_auth_mode(&state).await;
+    clear_auth_mode(&state, &identity.user_id).await;
     Ok(Json(ActionResponse::ok("Auth cancelled")))
 }
 
 /// Clear pending auth mode on the active thread.
-pub async fn clear_auth_mode(state: &GatewayState) {
+pub async fn clear_auth_mode(state: &GatewayState, user_id: &str) {
     if let Some(ref sm) = state.session_manager {
-        let session = sm.get_or_create_session(&state.user_id).await;
+        let session = sm.get_or_create_session(user_id).await;
         let mut sess = session.lock().await;
         if let Some(thread_id) = sess.active_thread
             && let Some(thread) = sess.threads.get_mut(&thread_id)
@@ -226,8 +240,9 @@ pub async fn clear_auth_mode(state: &GatewayState) {
 
 pub async fn chat_events_handler(
     State(state): State<Arc<GatewayState>>,
+    AuthenticatedUser(user): AuthenticatedUser,
 ) -> Result<impl IntoResponse, (StatusCode, String)> {
-    state.sse.subscribe().ok_or((
+    state.sse.subscribe(Some(user.user_id)).ok_or((
         StatusCode::SERVICE_UNAVAILABLE,
         "Too many connections".to_string(),
     ))
@@ -237,6 +252,7 @@ pub async fn chat_ws_handler(
     headers: axum::http::HeaderMap,
     ws: WebSocketUpgrade,
     State(state): State<Arc<GatewayState>>,
+    AuthenticatedUser(identity): AuthenticatedUser,
 ) -> Result<impl IntoResponse, (StatusCode, String)> {
     // Validate Origin header to prevent cross-site WebSocket hijacking.
     let origin = headers
@@ -262,7 +278,9 @@ pub async fn chat_ws_handler(
             "WebSocket origin not allowed".to_string(),
         ));
     }
-    Ok(ws.on_upgrade(move |socket| crate::channels::web::ws::handle_ws_connection(socket, state)))
+    Ok(ws.on_upgrade(move |socket| {
+        crate::channels::web::ws::handle_ws_connection(socket, state, identity)
+    }))
 }
 
 #[derive(Deserialize)]
@@ -274,6 +292,7 @@ pub struct HistoryQuery {
 
 pub async fn chat_history_handler(
     State(state): State<Arc<GatewayState>>,
+    AuthenticatedUser(identity): AuthenticatedUser,
     Query(query): Query<HistoryQuery>,
 ) -> Result<Json<HistoryResponse>, (StatusCode, String)> {
     let session_manager = state.session_manager.as_ref().ok_or((
@@ -281,7 +300,9 @@ pub async fn chat_history_handler(
         "Session manager not available".to_string(),
     ))?;
 
-    let session = session_manager.get_or_create_session(&state.user_id).await;
+    let session = session_manager
+        .get_or_create_session(&identity.user_id)
+        .await;
 
     let limit = query.limit.unwrap_or(50);
     let before_cursor = query
@@ -314,7 +335,7 @@ pub async fn chat_history_handler(
         && let Some(ref store) = state.store
     {
         let owned = store
-            .conversation_belongs_to_user(thread_id, &state.user_id)
+            .conversation_belongs_to_user(thread_id, &identity.user_id)
             .await
             .unwrap_or(false);
         if !owned {
@@ -377,8 +398,10 @@ pub async fn chat_history_handler(
                                 truncate_preview(&s, 500)
                             }),
                             error: tc.error.clone(),
+                            rationale: tc.rationale.clone(),
                         })
                         .collect(),
+                    narrative: t.narrative.clone(),
                 })
                 .collect();
 
@@ -434,24 +457,27 @@ pub async fn chat_history_handler(
 
 pub async fn chat_threads_handler(
     State(state): State<Arc<GatewayState>>,
+    AuthenticatedUser(identity): AuthenticatedUser,
 ) -> Result<Json<ThreadListResponse>, (StatusCode, String)> {
     let session_manager = state.session_manager.as_ref().ok_or((
         StatusCode::SERVICE_UNAVAILABLE,
         "Session manager not available".to_string(),
     ))?;
 
-    let session = session_manager.get_or_create_session(&state.user_id).await;
+    let session = session_manager
+        .get_or_create_session(&identity.user_id)
+        .await;
 
     // Try DB first for persistent thread list
     if let Some(ref store) = state.store {
         // Auto-create assistant thread if it doesn't exist
         let assistant_id = store
-            .get_or_create_assistant_conversation(&state.user_id, "gateway")
+            .get_or_create_assistant_conversation(&identity.user_id, "gateway")
             .await
             .map_err(|e| (StatusCode::INTERNAL_SERVER_ERROR, e.to_string()))?;
 
         if let Ok(summaries) = store
-            .list_conversations_all_channels(&state.user_id, 50)
+            .list_conversations_all_channels(&identity.user_id, 50)
             .await
         {
             let mut assistant_thread = None;
@@ -534,13 +560,16 @@ pub async fn chat_threads_handler(
 
 pub async fn chat_new_thread_handler(
     State(state): State<Arc<GatewayState>>,
+    AuthenticatedUser(identity): AuthenticatedUser,
 ) -> Result<Json<ThreadInfo>, (StatusCode, String)> {
     let session_manager = state.session_manager.as_ref().ok_or((
         StatusCode::SERVICE_UNAVAILABLE,
         "Session manager not available".to_string(),
     ))?;
 
-    let session = session_manager.get_or_create_session(&state.user_id).await;
+    let session = session_manager
+        .get_or_create_session(&identity.user_id)
+        .await;
     let (thread_id, info) = {
         let mut sess = session.lock().await;
         let thread = sess.create_thread();
@@ -562,12 +591,12 @@ pub async fn chat_new_thread_handler(
     // so that the subsequent loadThreads() call from the frontend sees it.
     if let Some(ref store) = state.store {
         match store
-            .ensure_conversation(thread_id, "gateway", &state.user_id, None)
+            .ensure_conversation(thread_id, "gateway", &identity.user_id, None)
             .await
         {
             Ok(true) => {}
             Ok(false) => tracing::warn!(
-                user = %state.user_id,
+                user = %identity.user_id,
                 thread_id = %thread_id,
                 "Skipped persisting new thread due to ownership/channel conflict"
             ),

src/channels/web/handlers/extensions.rs

@@ -8,11 +8,13 @@ use axum::{
     http::StatusCode,
 };
 
+use crate::channels::web::auth::AuthenticatedUser;
 use crate::channels::web::server::GatewayState;
 use crate::channels::web::types::*;
 
 pub async fn extensions_list_handler(
     State(state): State<Arc<GatewayState>>,
+    AuthenticatedUser(user): AuthenticatedUser,
 ) -> Result<Json<ExtensionListResponse>, (StatusCode, String)> {
     let ext_mgr = state.extension_manager.as_ref().ok_or((
         StatusCode::NOT_IMPLEMENTED,
@@ -20,7 +22,7 @@ pub async fn extensions_list_handler(
     ))?;
 
     let installed = ext_mgr
-        .list(None, false)
+        .list(None, false, &user.user_id)
         .await
         .map_err(|e| (StatusCode::INTERNAL_SERVER_ERROR, e.to_string()))?;
 
@@ -80,6 +82,7 @@ pub async fn extensions_list_handler(
 
 pub async fn extensions_tools_handler(
     State(state): State<Arc<GatewayState>>,
+    AuthenticatedUser(_user): AuthenticatedUser,
 ) -> Result<Json<ToolListResponse>, (StatusCode, String)> {
     let registry = state.tool_registry.as_ref().ok_or((
         StatusCode::SERVICE_UNAVAILABLE,
@@ -100,6 +103,7 @@ pub async fn extensions_tools_handler(
 
 pub async fn extensions_install_handler(
     State(state): State<Arc<GatewayState>>,
+    AuthenticatedUser(user): AuthenticatedUser,
     Json(req): Json<InstallExtensionRequest>,
 ) -> Result<Json<ActionResponse>, (StatusCode, String)> {
     let ext_mgr = state.extension_manager.as_ref().ok_or((
@@ -116,7 +120,7 @@ pub async fn extensions_install_handler(
     });
 
     match ext_mgr
-        .install(&req.name, req.url.as_deref(), kind_hint)
+        .install(&req.name, req.url.as_deref(), kind_hint, &user.user_id)
         .await
     {
         Ok(result) => Ok(Json(ActionResponse::ok(result.message))),
@@ -126,6 +130,7 @@ pub async fn extensions_install_handler(
 
 pub async fn extensions_remove_handler(
     State(state): State<Arc<GatewayState>>,
+    AuthenticatedUser(user): AuthenticatedUser,
     Path(name): Path<String>,
 ) -> Result<Json<ActionResponse>, (StatusCode, String)> {
     let ext_mgr = state.extension_manager.as_ref().ok_or((
@@ -133,7 +138,7 @@ pub async fn extensions_remove_handler(
         "Extension manager not available (secrets store required)".to_string(),
     ))?;
 
-    match ext_mgr.remove(&name).await {
+    match ext_mgr.remove(&name, &user.user_id).await {
         Ok(message) => Ok(Json(ActionResponse::ok(message))),
         Err(e) => Ok(Json(ActionResponse::fail(e.to_string()))),
     }

src/channels/web/handlers/jobs.rs

@@ -11,11 +11,13 @@ use axum::{
 use serde::Deserialize;
 use uuid::Uuid;
 
+use crate::channels::web::auth::AuthenticatedUser;
 use crate::channels::web::server::GatewayState;
 use crate::channels::web::types::*;
 
 pub async fn jobs_list_handler(
     State(state): State<Arc<GatewayState>>,
+    AuthenticatedUser(user): AuthenticatedUser,
 ) -> Result<Json<JobListResponse>, (StatusCode, String)> {
     let store = state.store.as_ref().ok_or((
         StatusCode::SERVICE_UNAVAILABLE,
@@ -25,8 +27,8 @@ pub async fn jobs_list_handler(
     let mut jobs: Vec<JobInfo> = Vec::new();
     let mut seen_ids: HashSet<Uuid> = HashSet::new();
 
-    // Fetch sandbox jobs from database.
+    // Fetch sandbox jobs scoped to this user.
-    match store.list_sandbox_jobs().await {
+    match store.list_sandbox_jobs_for_user(&user.user_id).await {
         Ok(sandbox_jobs) => {
             for j in &sandbox_jobs {
                 let ui_state = match j.status.as_str() {
@@ -50,8 +52,8 @@ pub async fn jobs_list_handler(
         }
     }
 
-    // Fetch agent (non-sandbox) jobs from database, deduplicating by ID.
+    // Fetch agent (non-sandbox) jobs scoped to this user, deduplicating by ID.
-    match store.list_agent_jobs().await {
+    match store.list_agent_jobs_for_user(&user.user_id).await {
         Ok(agent_jobs) => {
             for j in &agent_jobs {
                 if seen_ids.contains(&j.id) {
@@ -80,6 +82,7 @@ pub async fn jobs_list_handler(
 
 pub async fn jobs_summary_handler(
     State(state): State<Arc<GatewayState>>,
+    AuthenticatedUser(user): AuthenticatedUser,
 ) -> Result<Json<JobSummaryResponse>, (StatusCode, String)> {
     let store = state.store.as_ref().ok_or((
         StatusCode::SERVICE_UNAVAILABLE,
@@ -93,8 +96,8 @@ pub async fn jobs_summary_handler(
     let mut failed = 0;
     let mut stuck = 0;
 
-    // Sandbox job counts.
+    // Sandbox job counts scoped to this user.
-    match store.sandbox_job_summary().await {
+    match store.sandbox_job_summary_for_user(&user.user_id).await {
         Ok(s) => {
             total += s.total;
             pending += s.creating;
@@ -107,8 +110,8 @@ pub async fn jobs_summary_handler(
         }
     }
 
-    // Agent job counts.
+    // Agent job counts scoped to this user.
-    match store.agent_job_summary().await {
+    match store.agent_job_summary_for_user(&user.user_id).await {
         Ok(s) => {
             total += s.total;
             pending += s.pending;
@@ -134,6 +137,7 @@ pub async fn jobs_summary_handler(
 
 pub async fn jobs_detail_handler(
     State(state): State<Arc<GatewayState>>,
+    AuthenticatedUser(user): AuthenticatedUser,
     Path(id): Path<String>,
 ) -> Result<Json<JobDetailResponse>, (StatusCode, String)> {
     let store = state.store.as_ref().ok_or((
@@ -145,169 +149,213 @@ pub async fn jobs_detail_handler(
         .map_err(|_| (StatusCode::BAD_REQUEST, "Invalid job ID".to_string()))?;
 
     // Try sandbox job from DB first.
-    if let Ok(Some(job)) = store.get_sandbox_job(job_id).await {
+    match store.get_sandbox_job(job_id).await {
-        let browse_id = std::path::Path::new(&job.project_dir)
+        Ok(Some(job)) => {
-            .file_name()
+            if job.user_id != user.user_id {
-            .map(|n| n.to_string_lossy().to_string())
+                return Err((StatusCode::NOT_FOUND, "Job not found".to_string()));
-            .unwrap_or_else(|| job.id.to_string());
+            }
+            let browse_id = std::path::Path::new(&job.project_dir)
+                .file_name()
+                .map(|n| n.to_string_lossy().to_string())
+                .unwrap_or_else(|| job.id.to_string());
 
-        let ui_state = match job.status.as_str() {
+            let ui_state = match job.status.as_str() {
-            "creating" => "pending",
+                "creating" => "pending",
-            "running" => "in_progress",
+                "running" => "in_progress",
-            s => s,
+                s => s,
-        };
+            };
 
-        let elapsed_secs = job.started_at.map(|start| {
+            let elapsed_secs = job.started_at.map(|start| {
-            let end = job.completed_at.unwrap_or_else(chrono::Utc::now);
+                let end = job.completed_at.unwrap_or_else(chrono::Utc::now);
-            (end - start).num_seconds().max(0) as u64
+                (end - start).num_seconds().max(0) as u64
-        });
-
-        // Synthesize transitions from timestamps.
-        let mut transitions = Vec::new();
-        if let Some(started) = job.started_at {
-            transitions.push(TransitionInfo {
-                from: "creating".to_string(),
-                to: "running".to_string(),
-                timestamp: started.to_rfc3339(),
-                reason: None,
             });
-        }
-        if let Some(completed) = job.completed_at {
-            transitions.push(TransitionInfo {
-                from: "running".to_string(),
-                to: job.status.clone(),
-                timestamp: completed.to_rfc3339(),
-                reason: job.failure_reason.clone(),
-            });
-        }
 
-        let mode = store.get_sandbox_job_mode(job.id).await.ok().flatten();
+            // Synthesize transitions from timestamps.
-        let is_claude_code = mode.as_deref() == Some("claude_code");
+            let mut transitions = Vec::new();
+            if let Some(started) = job.started_at {
+                transitions.push(TransitionInfo {
+                    from: "creating".to_string(),
+                    to: "running".to_string(),
+                    timestamp: started.to_rfc3339(),
+                    reason: None,
+                });
+            }
+            if let Some(completed) = job.completed_at {
+                transitions.push(TransitionInfo {
+                    from: "running".to_string(),
+                    to: job.status.clone(),
+                    timestamp: completed.to_rfc3339(),
+                    reason: job.failure_reason.clone(),
+                });
+            }
 
-        return Ok(Json(JobDetailResponse {
+            let mode = store.get_sandbox_job_mode(job.id).await.ok().flatten();
-            id: job.id,
+            let is_claude_code = mode.as_deref() == Some("claude_code");
-            title: job.task.clone(),
+
-            description: String::new(),
+            return Ok(Json(JobDetailResponse {
-            state: ui_state.to_string(),
+                id: job.id,
-            user_id: job.user_id.clone(),
+                title: job.task.clone(),
-            created_at: job.created_at.to_rfc3339(),
+                description: String::new(),
-            started_at: job.started_at.map(|dt| dt.to_rfc3339()),
+                state: ui_state.to_string(),
-            completed_at: job.completed_at.map(|dt| dt.to_rfc3339()),
+                user_id: job.user_id.clone(),
-            elapsed_secs,
+                created_at: job.created_at.to_rfc3339(),
-            project_dir: Some(job.project_dir.clone()),
+                started_at: job.started_at.map(|dt| dt.to_rfc3339()),
-            browse_url: Some(format!("/projects/{}/", browse_id)),
+                completed_at: job.completed_at.map(|dt| dt.to_rfc3339()),
-            job_mode: mode.filter(|m| m != "worker"),
+                elapsed_secs,
-            transitions,
+                project_dir: Some(job.project_dir.clone()),
-            can_restart: state.job_manager.is_some(),
+                browse_url: Some(format!("/projects/{}/", browse_id)),
-            can_prompt: is_claude_code && state.prompt_queue.is_some(),
+                job_mode: mode.filter(|m| m != "worker"),
-            job_kind: Some("sandbox".to_string()),
+                transitions,
-        }));
+                can_restart: state.job_manager.is_some(),
+                can_prompt: is_claude_code && state.prompt_queue.is_some(),
+                job_kind: Some("sandbox".to_string()),
+            }));
+        }
+        Ok(None) => {}
+        Err(e) => {
+            return Err((
+                StatusCode::INTERNAL_SERVER_ERROR,
+                format!("Database error: {}", e),
+            ));
+        }
     }
 
     // Fall back to agent job from DB.
-    if let Ok(Some(ctx)) = store.get_job(job_id).await {
+    match store.get_job(job_id).await {
-        let elapsed_secs = ctx.started_at.map(|start| {
+        Ok(Some(ctx)) => {
-            let end = ctx.completed_at.unwrap_or_else(chrono::Utc::now);
+            if ctx.user_id != user.user_id {
-            (end - start).num_seconds().max(0) as u64
+                return Err((StatusCode::NOT_FOUND, "Job not found".to_string()));
-        });
+            }
+            let elapsed_secs = ctx.started_at.map(|start| {
+                let end = ctx.completed_at.unwrap_or_else(chrono::Utc::now);
+                (end - start).num_seconds().max(0) as u64
+            });
 
-        // Only show prompt bar for jobs that have a running worker (Pending/InProgress).
+            // Only show prompt bar for jobs that have a running worker (Pending/InProgress).
-        // Stuck jobs have no active worker loop, so messages would be silently dropped.
+            // Stuck jobs have no active worker loop, so messages would be silently dropped.
-        let is_promptable = matches!(
+            let is_promptable = matches!(
-            ctx.state,
+                ctx.state,
-            crate::context::JobState::Pending | crate::context::JobState::InProgress
+                crate::context::JobState::Pending | crate::context::JobState::InProgress
-        );
+            );
-        return Ok(Json(JobDetailResponse {
+            Ok(Json(JobDetailResponse {
-            id: ctx.job_id,
+                id: ctx.job_id,
-            title: ctx.title.clone(),
+                title: ctx.title.clone(),
-            description: ctx.description.clone(),
+                description: ctx.description.clone(),
-            state: ctx.state.to_string(),
+                state: ctx.state.to_string(),
-            user_id: ctx.user_id.clone(),
+                user_id: ctx.user_id.clone(),
-            created_at: ctx.created_at.to_rfc3339(),
+                created_at: ctx.created_at.to_rfc3339(),
-            started_at: ctx.started_at.map(|dt| dt.to_rfc3339()),
+                started_at: ctx.started_at.map(|dt| dt.to_rfc3339()),
-            completed_at: ctx.completed_at.map(|dt| dt.to_rfc3339()),
+                completed_at: ctx.completed_at.map(|dt| dt.to_rfc3339()),
-            elapsed_secs,
+                elapsed_secs,
-            project_dir: None,
+                project_dir: None,
-            browse_url: None,
+                browse_url: None,
-            job_mode: None,
+                job_mode: None,
-            transitions: Vec::new(),
+                transitions: Vec::new(),
-            can_restart: state.scheduler.is_some(),
+                can_restart: state.scheduler.is_some(),
-            can_prompt: is_promptable && state.scheduler.is_some(),
+                can_prompt: is_promptable && state.scheduler.is_some(),
-            job_kind: Some("agent".to_string()),
+                job_kind: Some("agent".to_string()),
-        }));
+            }))
+        }
+        Ok(None) => Err((StatusCode::NOT_FOUND, "Job not found".to_string())),
+        Err(e) => Err((
+            StatusCode::INTERNAL_SERVER_ERROR,
+            format!("Database error: {}", e),
+        )),
     }
-
-    Err((StatusCode::NOT_FOUND, "Job not found".to_string()))
 }
 
 pub async fn jobs_cancel_handler(
     State(state): State<Arc<GatewayState>>,
+    AuthenticatedUser(user): AuthenticatedUser,
     Path(id): Path<String>,
 ) -> Result<Json<serde_json::Value>, (StatusCode, String)> {
     let job_id = Uuid::parse_str(&id)
         .map_err(|_| (StatusCode::BAD_REQUEST, "Invalid job ID".to_string()))?;
 
     // Try sandbox job cancellation.
-    if let Some(ref store) = state.store
+    if let Some(ref store) = state.store {
-        && let Ok(Some(job)) = store.get_sandbox_job(job_id).await
+        match store.get_sandbox_job(job_id).await {
-    {
+            Ok(Some(job)) => {
-        if job.status == "running" || job.status == "creating" {
+                if job.user_id != user.user_id {
-            // Stop the container if we have a job manager.
+                    return Err((StatusCode::NOT_FOUND, "Job not found".to_string()));
-            if let Some(ref jm) = state.job_manager
+                }
-                && let Err(e) = jm.stop_job(job_id).await
+                if job.status == "running" || job.status == "creating" {
-            {
+                    if let Some(ref jm) = state.job_manager
-                tracing::warn!(job_id = %job_id, error = %e, "Failed to stop container during cancellation");
+                        && let Err(e) = jm.stop_job(job_id).await
+                    {
+                        tracing::warn!(job_id = %job_id, error = %e, "Failed to stop container during cancellation");
+                    }
+                    store
+                        .update_sandbox_job_status(
+                            job_id,
+                            "failed",
+                            Some(false),
+                            Some("Cancelled by user"),
+                            None,
+                            Some(chrono::Utc::now()),
+                        )
+                        .await
+                        .map_err(|e| (StatusCode::INTERNAL_SERVER_ERROR, e.to_string()))?;
+                }
+                return Ok(Json(serde_json::json!({
+                    "status": "cancelled",
+                    "job_id": job_id,
+                })));
+            }
+            Ok(None) => {}
+            Err(e) => {
+                return Err((
+                    StatusCode::INTERNAL_SERVER_ERROR,
+                    format!("Database error: {}", e),
+                ));
             }
-            store
-                .update_sandbox_job_status(
-                    job_id,
-                    "failed",
-                    Some(false),
-                    Some("Cancelled by user"),
-                    None,
-                    Some(chrono::Utc::now()),
-                )
-                .await
-                .map_err(|e| (StatusCode::INTERNAL_SERVER_ERROR, e.to_string()))?;
         }
-        return Ok(Json(serde_json::json!({
-            "status": "cancelled",
-            "job_id": job_id,
-        })));
     }
 
     // Fall back to agent job cancellation: stop the worker via the scheduler
     // (which updates the in-memory ContextManager AND aborts the task handle),
     // then persist the status to the DB as a fallback.
-    if let Some(ref store) = state.store
+    if let Some(ref store) = state.store {
-        && let Ok(Some(job)) = store.get_job(job_id).await
+        match store.get_job(job_id).await {
-    {
+            Ok(Some(job)) => {
-        if job.state.is_active() {
+                if job.user_id != user.user_id {
-            // Try to stop via scheduler (aborts the worker task + updates
+                    return Err((StatusCode::NOT_FOUND, "Job not found".to_string()));
-            // in-memory ContextManager). This is best-effort — the job may
+                }
-            // not be in the scheduler map if it already finished.
+                if job.state.is_active() {
-            if let Some(ref slot) = state.scheduler
+                    // Try to stop via scheduler (aborts the worker task + updates
-                && let Some(ref scheduler) = *slot.read().await
+                    // in-memory ContextManager). This is best-effort — the job may
-            {
+                    // not be in the scheduler map if it already finished.
-                let _ = scheduler.stop(job_id).await;
+                    if let Some(ref slot) = state.scheduler
-            }
+                        && let Some(ref scheduler) = *slot.read().await
+                    {
+                        let _ = scheduler.stop(job_id).await;
+                    }
 
-            // Always persist cancellation to the DB so the state is
+                    // Always persist cancellation to the DB so the state is
-            // consistent even if the scheduler wasn't available or the
+                    // consistent even if the scheduler wasn't available or the
-            // job wasn't in its in-memory map.
+                    // job wasn't in its in-memory map.
-            store
+                    store
-                .update_job_status(
+                        .update_job_status(
-                    job_id,
+                            job_id,
-                    crate::context::JobState::Cancelled,
+                            crate::context::JobState::Cancelled,
-                    Some("Cancelled by user"),
+                            Some("Cancelled by user"),
-                )
+                        )
-                .await
+                        .await
-                .map_err(|e| (StatusCode::INTERNAL_SERVER_ERROR, e.to_string()))?;
+                        .map_err(|e| (StatusCode::INTERNAL_SERVER_ERROR, e.to_string()))?;
+                }
+                return Ok(Json(serde_json::json!({
+                    "status": "cancelled",
+                    "job_id": job_id,
+                })));
+            }
+            Ok(None) => {}
+            Err(e) => {
+                return Err((
+                    StatusCode::INTERNAL_SERVER_ERROR,
+                    format!("Database error: {}", e),
+                ));
+            }
         }
-        return Ok(Json(serde_json::json!({
-            "status": "cancelled",
-            "job_id": job_id,
-        })));
     }
 
     Err((StatusCode::NOT_FOUND, "Job not found".to_string()))
@@ -315,6 +363,7 @@ pub async fn jobs_cancel_handler(
 
 pub async fn jobs_restart_handler(
     State(state): State<Arc<GatewayState>>,
+    AuthenticatedUser(user): AuthenticatedUser,
     Path(id): Path<String>,
 ) -> Result<Json<serde_json::Value>, (StatusCode, String)> {
     let store = state.store.as_ref().ok_or((
@@ -326,146 +375,166 @@ pub async fn jobs_restart_handler(
         .map_err(|_| (StatusCode::BAD_REQUEST, "Invalid job ID".to_string()))?;
 
     // Try sandbox job restart first.
-    if let Ok(Some(old_job)) = store.get_sandbox_job(old_job_id).await {
+    match store.get_sandbox_job(old_job_id).await {
-        if old_job.status != "interrupted" && old_job.status != "failed" {
+        Ok(Some(old_job)) => {
+            if old_job.user_id != user.user_id {
+                return Err((StatusCode::NOT_FOUND, "Job not found".to_string()));
+            }
+            if old_job.status != "interrupted" && old_job.status != "failed" {
+                return Err((
+                    StatusCode::CONFLICT,
+                    format!("Cannot restart job in state '{}'", old_job.status),
+                ));
+            }
+
+            let jm = state.job_manager.as_ref().ok_or((
+                StatusCode::SERVICE_UNAVAILABLE,
+                "Sandbox not enabled".to_string(),
+            ))?;
+
+            // Enrich the task with failure context.
+            let task = if let Some(ref reason) = old_job.failure_reason {
+                format!(
+                    "Previous attempt failed: {}. Retry: {}",
+                    reason, old_job.task
+                )
+            } else {
+                old_job.task.clone()
+            };
+
+            let new_job_id = Uuid::new_v4();
+            let now = chrono::Utc::now();
+
+            let record = crate::history::SandboxJobRecord {
+                id: new_job_id,
+                task: task.clone(),
+                status: "creating".to_string(),
+                user_id: old_job.user_id.clone(),
+                project_dir: old_job.project_dir.clone(),
+                success: None,
+                failure_reason: None,
+                created_at: now,
+                started_at: None,
+                completed_at: None,
+                credential_grants_json: old_job.credential_grants_json.clone(),
+            };
+            store
+                .save_sandbox_job(&record)
+                .await
+                .map_err(|e| (StatusCode::INTERNAL_SERVER_ERROR, e.to_string()))?;
+
+            let mode = match store.get_sandbox_job_mode(old_job_id).await {
+                Ok(Some(m)) if m == "claude_code" => {
+                    crate::orchestrator::job_manager::JobMode::ClaudeCode
+                }
+                _ => crate::orchestrator::job_manager::JobMode::Worker,
+            };
+
+            let credential_grants: Vec<crate::orchestrator::auth::CredentialGrant> =
+                serde_json::from_str(&old_job.credential_grants_json).unwrap_or_else(|e| {
+                    tracing::warn!(
+                        job_id = %old_job.id,
+                        "Failed to deserialize credential grants from stored job: {}. \
+                         Restarted job will have no credentials.",
+                        e
+                    );
+                    vec![]
+                });
+
+            let project_dir = std::path::PathBuf::from(&old_job.project_dir);
+            let _token = jm
+                .create_job(
+                    new_job_id,
+                    &task,
+                    Some(project_dir),
+                    mode,
+                    credential_grants,
+                )
+                .await
+                .map_err(|e| {
+                    (
+                        StatusCode::INTERNAL_SERVER_ERROR,
+                        format!("Failed to create container: {}", e),
+                    )
+                })?;
+
+            store
+                .update_sandbox_job_status(new_job_id, "running", None, None, Some(now), None)
+                .await
+                .map_err(|e| (StatusCode::INTERNAL_SERVER_ERROR, e.to_string()))?;
+
+            return Ok(Json(serde_json::json!({
+                "status": "restarted",
+                "old_job_id": old_job_id,
+                "new_job_id": new_job_id,
+            })));
+        }
+        Ok(None) => {}
+        Err(e) => {
             return Err((
-                StatusCode::CONFLICT,
+                StatusCode::INTERNAL_SERVER_ERROR,
-                format!("Cannot restart job in state '{}'", old_job.status),
+                format!("Database error: {}", e),
             ));
         }
-
-        let jm = state.job_manager.as_ref().ok_or((
-            StatusCode::SERVICE_UNAVAILABLE,
-            "Sandbox not enabled".to_string(),
-        ))?;
-
-        // Enrich the task with failure context.
-        let task = if let Some(ref reason) = old_job.failure_reason {
-            format!(
-                "Previous attempt failed: {}. Retry: {}",
-                reason, old_job.task
-            )
-        } else {
-            old_job.task.clone()
-        };
-
-        let new_job_id = Uuid::new_v4();
-        let now = chrono::Utc::now();
-
-        let record = crate::history::SandboxJobRecord {
-            id: new_job_id,
-            task: task.clone(),
-            status: "creating".to_string(),
-            user_id: old_job.user_id.clone(),
-            project_dir: old_job.project_dir.clone(),
-            success: None,
-            failure_reason: None,
-            created_at: now,
-            started_at: None,
-            completed_at: None,
-            credential_grants_json: old_job.credential_grants_json.clone(),
-        };
-        store
-            .save_sandbox_job(&record)
-            .await
-            .map_err(|e| (StatusCode::INTERNAL_SERVER_ERROR, e.to_string()))?;
-
-        let mode = match store.get_sandbox_job_mode(old_job_id).await {
-            Ok(Some(m)) if m == "claude_code" => {
-                crate::orchestrator::job_manager::JobMode::ClaudeCode
-            }
-            _ => crate::orchestrator::job_manager::JobMode::Worker,
-        };
-
-        let credential_grants: Vec<crate::orchestrator::auth::CredentialGrant> =
-            serde_json::from_str(&old_job.credential_grants_json).unwrap_or_else(|e| {
-                tracing::warn!(
-                    job_id = %old_job.id,
-                    "Failed to deserialize credential grants from stored job: {}. \
-                     Restarted job will have no credentials.",
-                    e
-                );
-                vec![]
-            });
-
-        let project_dir = std::path::PathBuf::from(&old_job.project_dir);
-        let _token = jm
-            .create_job(
-                new_job_id,
-                &task,
-                Some(project_dir),
-                mode,
-                credential_grants,
-            )
-            .await
-            .map_err(|e| {
-                (
-                    StatusCode::INTERNAL_SERVER_ERROR,
-                    format!("Failed to create container: {}", e),
-                )
-            })?;
-
-        store
-            .update_sandbox_job_status(new_job_id, "running", None, None, Some(now), None)
-            .await
-            .map_err(|e| (StatusCode::INTERNAL_SERVER_ERROR, e.to_string()))?;
-
-        return Ok(Json(serde_json::json!({
-            "status": "restarted",
-            "old_job_id": old_job_id,
-            "new_job_id": new_job_id,
-        })));
     }
 
     // Try agent job restart: dispatch a new job via the scheduler.
-    if let Ok(Some(old_job)) = store.get_job(old_job_id).await {
+    match store.get_job(old_job_id).await {
-        if old_job.state.is_active() {
+        Ok(Some(old_job)) => {
-            return Err((
+            if old_job.user_id != user.user_id {
-                StatusCode::CONFLICT,
+                return Err((StatusCode::NOT_FOUND, "Job not found".to_string()));
-                format!("Cannot restart job in state '{}'", old_job.state),
+            }
-            ));
+            if old_job.state.is_active() {
+                return Err((
+                    StatusCode::CONFLICT,
+                    format!("Cannot restart job in state '{}'", old_job.state),
+                ));
+            }
+
+            let slot = state.scheduler.as_ref().ok_or((
+                StatusCode::SERVICE_UNAVAILABLE,
+                "Scheduler not available".to_string(),
+            ))?;
+            let scheduler_guard = slot.read().await;
+            let scheduler = scheduler_guard.as_ref().ok_or((
+                StatusCode::SERVICE_UNAVAILABLE,
+                "Agent not started yet".to_string(),
+            ))?;
+
+            // Look up failure reason (O(1) point lookup).
+            let failure_reason = store
+                .get_agent_job_failure_reason(old_job_id)
+                .await
+                .ok()
+                .flatten()
+                .unwrap_or_default();
+
+            let title = if !failure_reason.is_empty() {
+                format!(
+                    "Previous attempt failed: {}. Retry: {}",
+                    failure_reason, old_job.title
+                )
+            } else {
+                old_job.title.clone()
+            };
+
+            let new_job_id = scheduler
+                .dispatch_job(&old_job.user_id, &title, &old_job.description, None)
+                .await
+                .map_err(|e| (StatusCode::INTERNAL_SERVER_ERROR, e.to_string()))?;
+
+            Ok(Json(serde_json::json!({
+                "status": "restarted",
+                "old_job_id": old_job_id,
+                "new_job_id": new_job_id,
+            })))
         }
-
+        Ok(None) => Err((StatusCode::NOT_FOUND, "Job not found".to_string())),
-        let slot = state.scheduler.as_ref().ok_or((
+        Err(e) => Err((
-            StatusCode::SERVICE_UNAVAILABLE,
+            StatusCode::INTERNAL_SERVER_ERROR,
-            "Scheduler not available".to_string(),
+            format!("Database error: {}", e),
-        ))?;
+        )),
-        let scheduler_guard = slot.read().await;
-        let scheduler = scheduler_guard.as_ref().ok_or((
-            StatusCode::SERVICE_UNAVAILABLE,
-            "Agent not started yet".to_string(),
-        ))?;
-
-        // Look up failure reason (O(1) point lookup).
-        let failure_reason = store
-            .get_agent_job_failure_reason(old_job_id)
-            .await
-            .ok()
-            .flatten()
-            .unwrap_or_default();
-
-        let title = if !failure_reason.is_empty() {
-            format!(
-                "Previous attempt failed: {}. Retry: {}",
-                failure_reason, old_job.title
-            )
-        } else {
-            old_job.title.clone()
-        };
-
-        let new_job_id = scheduler
-            .dispatch_job(&old_job.user_id, &title, &old_job.description, None)
-            .await
-            .map_err(|e| (StatusCode::INTERNAL_SERVER_ERROR, e.to_string()))?;
-
-        return Ok(Json(serde_json::json!({
-            "status": "restarted",
-            "old_job_id": old_job_id,
-            "new_job_id": new_job_id,
-        })));
     }
-
-    Err((StatusCode::NOT_FOUND, "Job not found".to_string()))
 }
 
 /// Submit a follow-up prompt to a running job.
@@ -476,6 +545,7 @@ pub async fn jobs_restart_handler(
 /// - Worker-mode sandbox jobs → not supported (no mechanism to inject)
 pub async fn jobs_prompt_handler(
     State(state): State<Arc<GatewayState>>,
+    AuthenticatedUser(user): AuthenticatedUser,
     Path(id): Path<String>,
     Json(body): Json<serde_json::Value>,
 ) -> Result<Json<serde_json::Value>, (StatusCode, String)> {
@@ -494,10 +564,15 @@ pub async fn jobs_prompt_handler(
 
     let done = body.get("done").and_then(|v| v.as_bool()).unwrap_or(false);
 
-    // Try sandbox job path: check if we have a sandbox record for this ID.
+    // Try sandbox job path first: verify ownership, then route to Claude Code or reject.
     if let Some(ref s) = state.store
-        && let Ok(Some(_)) = s.get_sandbox_job(job_id).await
+        && let Ok(Some(sandbox_job)) = s.get_sandbox_job(job_id).await
     {
+        // Verify ownership.
+        if sandbox_job.user_id != user.user_id {
+            return Err((StatusCode::NOT_FOUND, "Job not found".to_string()));
+        }
+
         // It's a sandbox job. Check if Claude Code mode.
         let mode = s.get_sandbox_job_mode(job_id).await.ok().flatten();
         if mode.as_deref() == Some("claude_code") {
@@ -522,7 +597,26 @@ pub async fn jobs_prompt_handler(
         }
     }
 
-    // Try agent job path: send via scheduler.
+    // Try agent job path: verify ownership, then send via scheduler.
+    if let Some(ref store) = state.store {
+        match store.get_job(job_id).await {
+            Ok(Some(agent_job)) => {
+                if agent_job.user_id != user.user_id {
+                    return Err((StatusCode::NOT_FOUND, "Job not found".to_string()));
+                }
+            }
+            Ok(None) => {
+                return Err((StatusCode::NOT_FOUND, "Job not found".to_string()));
+            }
+            Err(e) => {
+                return Err((
+                    StatusCode::INTERNAL_SERVER_ERROR,
+                    format!("Database error: {}", e),
+                ));
+            }
+        }
+    }
+
     let slot = state.scheduler.as_ref().ok_or((
         StatusCode::NOT_IMPLEMENTED,
         "Agent job prompts require the scheduler to be configured".to_string(),
@@ -550,6 +644,7 @@ pub async fn jobs_prompt_handler(
 /// Load persisted job events for a job (for history replay on page open).
 pub async fn jobs_events_handler(
     State(state): State<Arc<GatewayState>>,
+    AuthenticatedUser(user): AuthenticatedUser,
     Path(id): Path<String>,
 ) -> Result<Json<serde_json::Value>, (StatusCode, String)> {
     let store = state.store.as_ref().ok_or((
@@ -561,6 +656,24 @@ pub async fn jobs_events_handler(
         .parse()
         .map_err(|_| (StatusCode::BAD_REQUEST, "Invalid job ID".to_string()))?;
 
+    // Verify ownership before returning events.
+    match store.get_sandbox_job(job_id).await {
+        Ok(Some(job)) => {
+            if job.user_id != user.user_id {
+                return Err((StatusCode::NOT_FOUND, "Job not found".to_string()));
+            }
+        }
+        Ok(None) => {
+            return Err((StatusCode::NOT_FOUND, "Job not found".to_string()));
+        }
+        Err(e) => {
+            return Err((
+                StatusCode::INTERNAL_SERVER_ERROR,
+                format!("Database error: {}", e),
+            ));
+        }
+    }
+
     let events = store
         .list_job_events(job_id, None)
         .await
@@ -593,6 +706,7 @@ pub struct FilePathQuery {
 
 pub async fn job_files_list_handler(
     State(state): State<Arc<GatewayState>>,
+    AuthenticatedUser(user): AuthenticatedUser,
     Path(id): Path<String>,
     Query(query): Query<FilePathQuery>,
 ) -> Result<Json<ProjectFilesResponse>, (StatusCode, String)> {
@@ -610,6 +724,10 @@ pub async fn job_files_list_handler(
         .map_err(|e| (StatusCode::INTERNAL_SERVER_ERROR, e.to_string()))?
         .ok_or((StatusCode::NOT_FOUND, "Job not found".to_string()))?;
 
+    if job.user_id != user.user_id {
+        return Err((StatusCode::NOT_FOUND, "Job not found".to_string()));
+    }
+
     let base = std::path::PathBuf::from(&job.project_dir);
     let rel_path = query.path.as_deref().unwrap_or("");
     let target = base.join(rel_path);
@@ -656,6 +774,7 @@ pub async fn job_files_list_handler(
 
 pub async fn job_files_read_handler(
     State(state): State<Arc<GatewayState>>,
+    AuthenticatedUser(user): AuthenticatedUser,
     Path(id): Path<String>,
     Query(query): Query<FilePathQuery>,
 ) -> Result<Json<ProjectFileReadResponse>, (StatusCode, String)> {
@@ -673,6 +792,10 @@ pub async fn job_files_read_handler(
         .map_err(|e| (StatusCode::INTERNAL_SERVER_ERROR, e.to_string()))?
         .ok_or((StatusCode::NOT_FOUND, "Job not found".to_string()))?;
 
+    if job.user_id != user.user_id {
+        return Err((StatusCode::NOT_FOUND, "Job not found".to_string()));
+    }
+
     let path = query.path.as_deref().ok_or((
         StatusCode::BAD_REQUEST,
         "path parameter required".to_string(),

src/channels/web/handlers/memory.rs

@@ -9,8 +9,27 @@ use axum::{
 };
 use serde::Deserialize;
 
+use crate::channels::web::auth::{AuthenticatedUser, UserIdentity};
 use crate::channels::web::server::GatewayState;
 use crate::channels::web::types::*;
+use crate::workspace::Workspace;
+
+/// Resolve the workspace for the authenticated user.
+///
+/// Prefers `workspace_pool` (multi-user mode) when available, falling back
+/// to the single-user `state.workspace`.
+pub(crate) async fn resolve_workspace(
+    state: &GatewayState,
+    user: &UserIdentity,
+) -> Result<Arc<Workspace>, (StatusCode, String)> {
+    if let Some(ref pool) = state.workspace_pool {
+        return Ok(pool.get_or_create(user).await);
+    }
+    state.workspace.as_ref().cloned().ok_or((
+        StatusCode::SERVICE_UNAVAILABLE,
+        "Workspace not available".to_string(),
+    ))
+}
 
 #[derive(Deserialize)]
 pub struct TreeQuery {
@@ -20,12 +39,10 @@ pub struct TreeQuery {
 
 pub async fn memory_tree_handler(
     State(state): State<Arc<GatewayState>>,
+    AuthenticatedUser(user): AuthenticatedUser,
     Query(_query): Query<TreeQuery>,
 ) -> Result<Json<MemoryTreeResponse>, (StatusCode, String)> {
-    let workspace = state.workspace.as_ref().ok_or((
+    let workspace = resolve_workspace(&state, &user).await?;
-        StatusCode::SERVICE_UNAVAILABLE,
-        "Workspace not available".to_string(),
-    ))?;
 
     // Build tree from list_all (flat list of all paths)
     let all_paths = workspace
@@ -68,12 +85,10 @@ pub struct ListQuery {
 
 pub async fn memory_list_handler(
     State(state): State<Arc<GatewayState>>,
+    AuthenticatedUser(user): AuthenticatedUser,
     Query(query): Query<ListQuery>,
 ) -> Result<Json<MemoryListResponse>, (StatusCode, String)> {
-    let workspace = state.workspace.as_ref().ok_or((
+    let workspace = resolve_workspace(&state, &user).await?;
-        StatusCode::SERVICE_UNAVAILABLE,
-        "Workspace not available".to_string(),
-    ))?;
 
     let path = query.path.as_deref().unwrap_or("");
     let entries = workspace
@@ -104,12 +119,10 @@ pub struct ReadQuery {
 
 pub async fn memory_read_handler(
     State(state): State<Arc<GatewayState>>,
+    AuthenticatedUser(user): AuthenticatedUser,
     Query(query): Query<ReadQuery>,
 ) -> Result<Json<MemoryReadResponse>, (StatusCode, String)> {
-    let workspace = state.workspace.as_ref().ok_or((
+    let workspace = resolve_workspace(&state, &user).await?;
-        StatusCode::SERVICE_UNAVAILABLE,
-        "Workspace not available".to_string(),
-    ))?;
 
     let doc = workspace
         .read(&query.path)
@@ -123,17 +136,75 @@ pub async fn memory_read_handler(
     }))
 }
 
-// memory_write_handler lives in server.rs (layer-aware version with append,
+pub async fn memory_write_handler(
-// privacy redirect, and proper error status codes).
+    State(state): State<Arc<GatewayState>>,
+    AuthenticatedUser(user): AuthenticatedUser,
+    Json(req): Json<MemoryWriteRequest>,
+) -> Result<Json<MemoryWriteResponse>, (StatusCode, String)> {
+    let workspace = resolve_workspace(&state, &user).await?;
+
+    // Route through layer-aware methods when a layer is specified.
+    //
+    // Note: unlike MemoryWriteTool, this endpoint does NOT block writes to
+    // identity files (IDENTITY.md, SOUL.md, etc.). The HTTP API is an
+    // authenticated admin interface; the supervisor uses it to seed identity
+    // files at startup. Identity-file protection is enforced at the tool
+    // layer (LLM-facing) where the write originates from an untrusted agent.
+    if let Some(ref layer_name) = req.layer {
+        let result = if req.append {
+            workspace
+                .append_to_layer(layer_name, &req.path, &req.content, req.force)
+                .await
+        } else {
+            workspace
+                .write_to_layer(layer_name, &req.path, &req.content, req.force)
+                .await
+        }
+        .map_err(|e| {
+            use crate::error::WorkspaceError;
+            let status = match &e {
+                WorkspaceError::LayerNotFound { .. } => StatusCode::BAD_REQUEST,
+                WorkspaceError::LayerReadOnly { .. } => StatusCode::FORBIDDEN,
+                WorkspaceError::PrivacyRedirectFailed => StatusCode::UNPROCESSABLE_ENTITY,
+                _ => StatusCode::INTERNAL_SERVER_ERROR,
+            };
+            (status, e.to_string())
+        })?;
+        return Ok(Json(MemoryWriteResponse {
+            path: req.path,
+            status: "written",
+            redirected: Some(result.redirected),
+            actual_layer: Some(result.actual_layer),
+        }));
+    }
+
+    // Non-layer path: honor the append field
+    if req.append {
+        workspace
+            .append(&req.path, &req.content)
+            .await
+            .map_err(|e| (StatusCode::INTERNAL_SERVER_ERROR, e.to_string()))?;
+    } else {
+        workspace
+            .write(&req.path, &req.content)
+            .await
+            .map_err(|e| (StatusCode::INTERNAL_SERVER_ERROR, e.to_string()))?;
+    }
+
+    Ok(Json(MemoryWriteResponse {
+        path: req.path,
+        status: "written",
+        redirected: None,
+        actual_layer: None,
+    }))
+}
 
 pub async fn memory_search_handler(
     State(state): State<Arc<GatewayState>>,
+    AuthenticatedUser(user): AuthenticatedUser,
     Json(req): Json<MemorySearchRequest>,
 ) -> Result<Json<MemorySearchResponse>, (StatusCode, String)> {
-    let workspace = state.workspace.as_ref().ok_or((
+    let workspace = resolve_workspace(&state, &user).await?;
-        StatusCode::SERVICE_UNAVAILABLE,
-        "Workspace not available".to_string(),
-    ))?;
 
     let limit = req.limit.unwrap_or(10);
     let results = workspace
@@ -142,10 +213,10 @@ pub async fn memory_search_handler(
         .map_err(|e| (StatusCode::INTERNAL_SERVER_ERROR, e.to_string()))?;
 
     let hits: Vec<SearchHit> = results
-        .into_iter()
+        .iter()
         .map(|r| SearchHit {
-            path: r.document_path,
+            path: r.document_id.to_string(),
-            content: r.content,
+            content: r.content.clone(),
             score: r.score as f64,
         })
         .collect();

src/channels/web/handlers/mod.rs

@@ -1,13 +1,10 @@
 //! Handler modules for the web gateway API.
 //!
 //! Each module groups related endpoint handlers by domain.
-//!
-//! # Migration status
-//!
-//! `skills` is the canonical implementation used by `server.rs`.
-//! The remaining modules are in-progress migrations from inline server.rs
-//! handlers; their functions are not yet wired up, hence the `dead_code` allow.
 
+pub mod jobs;
+pub mod memory;
+pub mod routines;
 pub mod skills;
 
 // Modules not yet wired into server.rs router -- suppress dead_code until
@@ -17,12 +14,6 @@ pub mod chat;
 #[allow(dead_code)]
 pub mod extensions;
 #[allow(dead_code)]
-pub mod jobs;
-#[allow(dead_code)]
-pub mod memory;
-#[allow(dead_code)]
-pub mod routines;
-#[allow(dead_code)]
 pub mod settings;
 #[allow(dead_code)]
 pub mod static_files;

src/channels/web/handlers/routines.rs

@@ -11,12 +11,14 @@ use serde::Deserialize;
 use uuid::Uuid;
 
 use crate::agent::routine::{Trigger, next_cron_fire};
+use crate::channels::web::auth::AuthenticatedUser;
 use crate::channels::web::server::GatewayState;
 use crate::channels::web::types::*;
 use crate::error::RoutineError;
 
 pub async fn routines_list_handler(
     State(state): State<Arc<GatewayState>>,
+    AuthenticatedUser(user): AuthenticatedUser,
 ) -> Result<Json<RoutineListResponse>, (StatusCode, String)> {
     let store = state.store.as_ref().ok_or((
         StatusCode::SERVICE_UNAVAILABLE,
@@ -24,7 +26,7 @@ pub async fn routines_list_handler(
     ))?;
 
     let routines = store
-        .list_all_routines()
+        .list_routines(&user.user_id)
         .await
         .map_err(|e| (StatusCode::INTERNAL_SERVER_ERROR, e.to_string()))?;
 
@@ -35,6 +37,7 @@ pub async fn routines_list_handler(
 
 pub async fn routines_summary_handler(
     State(state): State<Arc<GatewayState>>,
+    AuthenticatedUser(user): AuthenticatedUser,
 ) -> Result<Json<RoutineSummaryResponse>, (StatusCode, String)> {
     let store = state.store.as_ref().ok_or((
         StatusCode::SERVICE_UNAVAILABLE,
@@ -42,7 +45,7 @@ pub async fn routines_summary_handler(
     ))?;
 
     let routines = store
-        .list_all_routines()
+        .list_routines(&user.user_id)
         .await
         .map_err(|e| (StatusCode::INTERNAL_SERVER_ERROR, e.to_string()))?;
 
@@ -78,6 +81,7 @@ pub async fn routines_summary_handler(
 
 pub async fn routines_detail_handler(
     State(state): State<Arc<GatewayState>>,
+    AuthenticatedUser(user): AuthenticatedUser,
     Path(id): Path<String>,
 ) -> Result<Json<RoutineDetailResponse>, (StatusCode, String)> {
     let store = state.store.as_ref().ok_or((
@@ -94,6 +98,10 @@ pub async fn routines_detail_handler(
         .map_err(|e| (StatusCode::INTERNAL_SERVER_ERROR, e.to_string()))?
         .ok_or((StatusCode::NOT_FOUND, "Routine not found".to_string()))?;
 
+    if routine.user_id != user.user_id {
+        return Err((StatusCode::NOT_FOUND, "Routine not found".to_string()));
+    }
+
     let runs = store
         .list_routine_runs(routine_id, 20)
         .await
@@ -106,7 +114,7 @@ pub async fn routines_detail_handler(
             trigger_type: run.trigger_type.clone(),
             started_at: run.started_at.to_rfc3339(),
             completed_at: run.completed_at.map(|dt| dt.to_rfc3339()),
-            status: format!("{:?}", run.status),
+            status: run.status.to_string(),
             result_summary: run.result_summary.clone(),
             tokens_used: run.tokens_used,
             job_id: run.job_id,
@@ -137,6 +145,7 @@ pub async fn routines_detail_handler(
 
 pub async fn routines_trigger_handler(
     State(state): State<Arc<GatewayState>>,
+    AuthenticatedUser(user): AuthenticatedUser,
     Path(id): Path<String>,
 ) -> Result<Json<serde_json::Value>, (StatusCode, String)> {
     // Clone the Arc out of the lock to avoid holding the RwLock across .await.
@@ -152,7 +161,7 @@ pub async fn routines_trigger_handler(
         .map_err(|_| (StatusCode::BAD_REQUEST, "Invalid routine ID".to_string()))?;
 
     let run_id = engine
-        .fire_manual(routine_id, Some(&state.user_id))
+        .fire_manual(routine_id, Some(&user.user_id))
         .await
         .map_err(|e| (routine_error_status(&e), e.to_string()))?;
 
@@ -170,6 +179,7 @@ pub struct ToggleRequest {
 
 pub async fn routines_toggle_handler(
     State(state): State<Arc<GatewayState>>,
+    AuthenticatedUser(user): AuthenticatedUser,
     Path(id): Path<String>,
     body: Option<Json<ToggleRequest>>,
 ) -> Result<Json<serde_json::Value>, (StatusCode, String)> {
@@ -187,6 +197,10 @@ pub async fn routines_toggle_handler(
         .map_err(|e| (StatusCode::INTERNAL_SERVER_ERROR, e.to_string()))?
         .ok_or((StatusCode::NOT_FOUND, "Routine not found".to_string()))?;
 
+    if routine.user_id != user.user_id {
+        return Err((StatusCode::NOT_FOUND, "Routine not found".to_string()));
+    }
+
     let was_enabled = routine.enabled;
     // If a specific value was provided, use it; otherwise toggle.
     routine.enabled = match body {
@@ -230,6 +244,7 @@ pub async fn routines_toggle_handler(
 
 pub async fn routines_delete_handler(
     State(state): State<Arc<GatewayState>>,
+    AuthenticatedUser(user): AuthenticatedUser,
     Path(id): Path<String>,
 ) -> Result<Json<serde_json::Value>, (StatusCode, String)> {
     let store = state.store.as_ref().ok_or((
@@ -240,6 +255,17 @@ pub async fn routines_delete_handler(
     let routine_id = Uuid::parse_str(&id)
         .map_err(|_| (StatusCode::BAD_REQUEST, "Invalid routine ID".to_string()))?;
 
+    // Verify ownership before deleting.
+    let routine = store
+        .get_routine(routine_id)
+        .await
+        .map_err(|e| (StatusCode::INTERNAL_SERVER_ERROR, e.to_string()))?
+        .ok_or((StatusCode::NOT_FOUND, "Routine not found".to_string()))?;
+
+    if routine.user_id != user.user_id {
+        return Err((StatusCode::NOT_FOUND, "Routine not found".to_string()));
+    }
+
     let deleted = store
         .delete_routine(routine_id)
         .await
@@ -261,8 +287,10 @@ pub async fn routines_delete_handler(
     }
 }
 
+#[allow(dead_code)] // Used by server.rs inline version; kept in sync here for future migration.
 pub async fn routines_runs_handler(
     State(state): State<Arc<GatewayState>>,
+    AuthenticatedUser(user): AuthenticatedUser,
     Path(id): Path<String>,
 ) -> Result<Json<serde_json::Value>, (StatusCode, String)> {
     let store = state.store.as_ref().ok_or((
@@ -273,6 +301,17 @@ pub async fn routines_runs_handler(
     let routine_id = Uuid::parse_str(&id)
         .map_err(|_| (StatusCode::BAD_REQUEST, "Invalid routine ID".to_string()))?;
 
+    // Verify ownership before listing runs.
+    let routine = store
+        .get_routine(routine_id)
+        .await
+        .map_err(|e| (StatusCode::INTERNAL_SERVER_ERROR, e.to_string()))?
+        .ok_or((StatusCode::NOT_FOUND, "Routine not found".to_string()))?;
+
+    if routine.user_id != user.user_id {
+        return Err((StatusCode::NOT_FOUND, "Routine not found".to_string()));
+    }
+
     let runs = store
         .list_routine_runs(routine_id, 50)
         .await
@@ -285,7 +324,7 @@ pub async fn routines_runs_handler(
             trigger_type: run.trigger_type.clone(),
             started_at: run.started_at.to_rfc3339(),
             completed_at: run.completed_at.map(|dt| dt.to_rfc3339()),
-            status: format!("{:?}", run.status),
+            status: run.status.to_string(),
             result_summary: run.result_summary.clone(),
             tokens_used: run.tokens_used,
             job_id: run.job_id,

src/channels/web/handlers/settings.rs

@@ -8,17 +8,19 @@ use axum::{
     http::StatusCode,
 };
 
+use crate::channels::web::auth::AuthenticatedUser;
 use crate::channels::web::server::GatewayState;
 use crate::channels::web::types::*;
 
 pub async fn settings_list_handler(
     State(state): State<Arc<GatewayState>>,
+    AuthenticatedUser(user): AuthenticatedUser,
 ) -> Result<Json<SettingsListResponse>, StatusCode> {
     let store = state
         .store
         .as_ref()
         .ok_or(StatusCode::SERVICE_UNAVAILABLE)?;
-    let rows = store.list_settings(&state.user_id).await.map_err(|e| {
+    let rows = store.list_settings(&user.user_id).await.map_err(|e| {
         tracing::error!("Failed to list settings: {}", e);
         StatusCode::INTERNAL_SERVER_ERROR
     })?;
@@ -37,6 +39,7 @@ pub async fn settings_list_handler(
 
 pub async fn settings_get_handler(
     State(state): State<Arc<GatewayState>>,
+    AuthenticatedUser(user): AuthenticatedUser,
     Path(key): Path<String>,
 ) -> Result<Json<SettingResponse>, StatusCode> {
     let store = state
@@ -44,7 +47,7 @@ pub async fn settings_get_handler(
         .as_ref()
         .ok_or(StatusCode::SERVICE_UNAVAILABLE)?;
     let row = store
-        .get_setting_full(&state.user_id, &key)
+        .get_setting_full(&user.user_id, &key)
         .await
         .map_err(|e| {
             tracing::error!("Failed to get setting '{}': {}", key, e);
@@ -61,6 +64,7 @@ pub async fn settings_get_handler(
 
 pub async fn settings_set_handler(
     State(state): State<Arc<GatewayState>>,
+    AuthenticatedUser(user): AuthenticatedUser,
     Path(key): Path<String>,
     Json(body): Json<SettingWriteRequest>,
 ) -> Result<StatusCode, StatusCode> {
@@ -69,7 +73,7 @@ pub async fn settings_set_handler(
         .as_ref()
         .ok_or(StatusCode::SERVICE_UNAVAILABLE)?;
     store
-        .set_setting(&state.user_id, &key, &body.value)
+        .set_setting(&user.user_id, &key, &body.value)
         .await
         .map_err(|e| {
             tracing::error!("Failed to set setting '{}': {}", key, e);
@@ -81,6 +85,7 @@ pub async fn settings_set_handler(
 
 pub async fn settings_delete_handler(
     State(state): State<Arc<GatewayState>>,
+    AuthenticatedUser(user): AuthenticatedUser,
     Path(key): Path<String>,
 ) -> Result<StatusCode, StatusCode> {
     let store = state
@@ -88,7 +93,7 @@ pub async fn settings_delete_handler(
         .as_ref()
         .ok_or(StatusCode::SERVICE_UNAVAILABLE)?;
     store
-        .delete_setting(&state.user_id, &key)
+        .delete_setting(&user.user_id, &key)
         .await
         .map_err(|e| {
             tracing::error!("Failed to delete setting '{}': {}", key, e);
@@ -100,12 +105,13 @@ pub async fn settings_delete_handler(
 
 pub async fn settings_export_handler(
     State(state): State<Arc<GatewayState>>,
+    AuthenticatedUser(user): AuthenticatedUser,
 ) -> Result<Json<SettingsExportResponse>, StatusCode> {
     let store = state
         .store
         .as_ref()
         .ok_or(StatusCode::SERVICE_UNAVAILABLE)?;
-    let settings = store.get_all_settings(&state.user_id).await.map_err(|e| {
+    let settings = store.get_all_settings(&user.user_id).await.map_err(|e| {
         tracing::error!("Failed to export settings: {}", e);
         StatusCode::INTERNAL_SERVER_ERROR
     })?;
@@ -115,6 +121,7 @@ pub async fn settings_export_handler(
 
 pub async fn settings_import_handler(
     State(state): State<Arc<GatewayState>>,
+    AuthenticatedUser(user): AuthenticatedUser,
     Json(body): Json<SettingsImportRequest>,
 ) -> Result<StatusCode, StatusCode> {
     let store = state
@@ -122,7 +129,7 @@ pub async fn settings_import_handler(
         .as_ref()
         .ok_or(StatusCode::SERVICE_UNAVAILABLE)?;
     store
-        .set_all_settings(&state.user_id, &body.settings)
+        .set_all_settings(&user.user_id, &body.settings)
         .await
         .map_err(|e| {
             tracing::error!("Failed to import settings: {}", e);

src/channels/web/handlers/skills.rs

@@ -8,11 +8,13 @@ use axum::{
     http::StatusCode,
 };
 
+use crate::channels::web::auth::AuthenticatedUser;
 use crate::channels::web::server::GatewayState;
 use crate::channels::web::types::*;
 
 pub async fn skills_list_handler(
     State(state): State<Arc<GatewayState>>,
+    AuthenticatedUser(_user): AuthenticatedUser,
 ) -> Result<Json<SkillListResponse>, (StatusCode, String)> {
     let registry = state.skill_registry.as_ref().ok_or((
         StatusCode::NOT_IMPLEMENTED,
@@ -45,6 +47,7 @@ pub async fn skills_list_handler(
 
 pub async fn skills_search_handler(
     State(state): State<Arc<GatewayState>>,
+    AuthenticatedUser(_user): AuthenticatedUser,
     Json(req): Json<SkillSearchRequest>,
 ) -> Result<Json<SkillSearchResponse>, (StatusCode, String)> {
     let registry = state.skill_registry.as_ref().ok_or((
@@ -119,6 +122,7 @@ pub async fn skills_search_handler(
 
 pub async fn skills_install_handler(
     State(state): State<Arc<GatewayState>>,
+    AuthenticatedUser(user): AuthenticatedUser,
     headers: axum::http::HeaderMap,
     Json(req): Json<SkillInstallRequest>,
 ) -> Result<Json<ActionResponse>, (StatusCode, String)> {
@@ -135,6 +139,8 @@ pub async fn skills_install_handler(
         ));
     }
 
+    tracing::info!(user_id = %user.user_id, skill = %req.name, "skill install requested");
+
     let registry = state.skill_registry.as_ref().ok_or((
         StatusCode::NOT_IMPLEMENTED,
         "Skills system not enabled".to_string(),
@@ -219,6 +225,7 @@ pub async fn skills_install_handler(
 
 pub async fn skills_remove_handler(
     State(state): State<Arc<GatewayState>>,
+    AuthenticatedUser(user): AuthenticatedUser,
     headers: axum::http::HeaderMap,
     Path(name): Path<String>,
 ) -> Result<Json<ActionResponse>, (StatusCode, String)> {
@@ -234,6 +241,8 @@ pub async fn skills_remove_handler(
         ));
     }
 
+    tracing::info!(user_id = %user.user_id, skill = %name, "skill remove requested");
+
     let registry = state.skill_registry.as_ref().ok_or((
         StatusCode::NOT_IMPLEMENTED,
         "Skills system not enabled".to_string(),

src/channels/web/handlers/static_files.rs

@@ -7,6 +7,7 @@ use axum::{
 };
 
 use crate::bootstrap::ironclaw_base_dir;
+use crate::channels::web::auth::AuthenticatedUser;
 use crate::channels::web::types::*;
 
 // --- Static file handlers ---
@@ -113,6 +114,7 @@ use crate::channels::web::server::GatewayState;
 
 pub async fn logs_events_handler(
     State(state): State<Arc<GatewayState>>,
+    AuthenticatedUser(_user): AuthenticatedUser,
 ) -> Result<
     Sse<impl futures::Stream<Item = Result<Event, Infallible>> + Send + 'static>,
     (StatusCode, String),
@@ -152,6 +154,7 @@ pub async fn logs_events_handler(
 
 pub async fn gateway_status_handler(
     State(state): State<Arc<GatewayState>>,
+    AuthenticatedUser(_user): AuthenticatedUser,
 ) -> Json<GatewayStatusResponse> {
     let sse_connections = state.sse.connection_count();
     let ws_connections = state

src/channels/web/mod.rs

@@ -31,6 +31,9 @@ pub mod ws;
 /// [`TestGatewayBuilder`](test_helpers::TestGatewayBuilder).
 pub mod test_helpers;
 
+#[cfg(test)]
+mod tests;
+
 use std::net::SocketAddr;
 use std::sync::Arc;
 
@@ -52,22 +55,24 @@ use crate::workspace::Workspace;
 
 use self::log_layer::{LogBroadcaster, LogLevelHandle};
 
+use self::auth::MultiAuthState;
 use self::server::GatewayState;
 use self::sse::SseManager;
-use self::types::SseEvent;
+use self::types::AppEvent;
 
 /// Web gateway channel implementing the Channel trait.
 pub struct GatewayChannel {
     config: GatewayConfig,
     state: Arc<GatewayState>,
-    /// The actual auth token in use (generated or from config).
+    /// Multi-user auth state (replaces bare auth_token).
-    auth_token: String,
+    auth: MultiAuthState,
 }
 
 impl GatewayChannel {
     /// Create a new gateway channel.
     ///
     /// If no auth token is configured, generates a random one and prints it.
+    /// Builds a single-user `MultiAuthState` from the config.
     pub fn new(config: GatewayConfig) -> Self {
         let auth_token = config.auth_token.clone().unwrap_or_else(|| {
             use rand::RngCore;
@@ -77,10 +82,13 @@ impl GatewayChannel {
             bytes.iter().map(|b| format!("{b:02x}")).collect()
         });
 
+        let auth = MultiAuthState::single(auth_token, config.user_id.clone());
+
         let state = Arc::new(GatewayState {
             msg_tx: tokio::sync::RwLock::new(None),
-            sse: SseManager::new(),
+            sse: Arc::new(SseManager::new()),
             workspace: None,
+            workspace_pool: None,
             session_manager: None,
             log_broadcaster: None,
             log_level_handle: None,
@@ -90,13 +98,14 @@ impl GatewayChannel {
             job_manager: None,
             prompt_queue: None,
             scheduler: None,
-            user_id: config.user_id.clone(),
+            owner_id: config.user_id.clone(),
+            default_sender_id: config.user_id.clone(),
             shutdown_tx: tokio::sync::RwLock::new(None),
             ws_tracker: Some(Arc::new(ws::WsConnectionTracker::new())),
             llm_provider: None,
             skill_registry: None,
             skill_catalog: None,
-            chat_rate_limiter: server::RateLimiter::new(30, 60),
+            chat_rate_limiter: server::PerUserRateLimiter::new(30, 60),
             oauth_rate_limiter: server::RateLimiter::new(10, 60),
             webhook_rate_limiter: server::RateLimiter::new(10, 60),
             registry_entries: Vec::new(),
@@ -109,7 +118,63 @@ impl GatewayChannel {
         Self {
             config,
             state,
-            auth_token,
+            auth,
+        }
+    }
+
+    /// Rebind the single-user auth identity to the durable owner scope while
+    /// preserving the configured gateway sender/routing identity.
+    pub fn with_owner_scope(mut self, owner_id: impl Into<String>) -> Self {
+        let owner_id = owner_id.into();
+        let single_user_token = if self.config.user_tokens.is_none() {
+            self.auth.first_token().map(ToOwned::to_owned)
+        } else {
+            None
+        };
+        if let Some(token) = single_user_token {
+            self.auth = MultiAuthState::single(token, owner_id.clone());
+        }
+        self.rebuild_state(|s| s.owner_id = owner_id);
+        self
+    }
+
+    /// Create a gateway channel with a pre-built multi-user auth state.
+    pub fn new_multi_auth(config: GatewayConfig, auth: MultiAuthState) -> Self {
+        let state = Arc::new(GatewayState {
+            msg_tx: tokio::sync::RwLock::new(None),
+            sse: Arc::new(SseManager::new()),
+            workspace: None,
+            workspace_pool: None,
+            session_manager: None,
+            log_broadcaster: None,
+            log_level_handle: None,
+            extension_manager: None,
+            tool_registry: None,
+            store: None,
+            job_manager: None,
+            prompt_queue: None,
+            scheduler: None,
+            owner_id: config.user_id.clone(),
+            default_sender_id: config.user_id.clone(),
+            shutdown_tx: tokio::sync::RwLock::new(None),
+            ws_tracker: Some(Arc::new(ws::WsConnectionTracker::new())),
+            llm_provider: None,
+            skill_registry: None,
+            skill_catalog: None,
+            chat_rate_limiter: server::PerUserRateLimiter::new(30, 60),
+            oauth_rate_limiter: server::RateLimiter::new(10, 60),
+            registry_entries: Vec::new(),
+            cost_guard: None,
+            routine_engine: Arc::new(tokio::sync::RwLock::new(None)),
+            startup_time: std::time::Instant::now(),
+            webhook_rate_limiter: server::RateLimiter::new(10, 60),
+            active_config: server::ActiveConfigSnapshot::default(),
+        });
+
+        Self {
+            config,
+            state,
+            auth,
         }
     }
 
@@ -118,8 +183,9 @@ impl GatewayChannel {
         let mut new_state = GatewayState {
             msg_tx: tokio::sync::RwLock::new(None),
             // Preserve the existing broadcast channel so sender handles remain valid.
-            sse: SseManager::from_sender(self.state.sse.sender()),
+            sse: Arc::new(SseManager::from_sender(self.state.sse.sender())),
             workspace: self.state.workspace.clone(),
+            workspace_pool: self.state.workspace_pool.clone(),
             session_manager: self.state.session_manager.clone(),
             log_broadcaster: self.state.log_broadcaster.clone(),
             log_level_handle: self.state.log_level_handle.clone(),
@@ -129,13 +195,14 @@ impl GatewayChannel {
             job_manager: self.state.job_manager.clone(),
             prompt_queue: self.state.prompt_queue.clone(),
             scheduler: self.state.scheduler.clone(),
-            user_id: self.state.user_id.clone(),
+            owner_id: self.state.owner_id.clone(),
+            default_sender_id: self.state.default_sender_id.clone(),
             shutdown_tx: tokio::sync::RwLock::new(None),
             ws_tracker: self.state.ws_tracker.clone(),
             llm_provider: self.state.llm_provider.clone(),
             skill_registry: self.state.skill_registry.clone(),
             skill_catalog: self.state.skill_catalog.clone(),
-            chat_rate_limiter: server::RateLimiter::new(30, 60),
+            chat_rate_limiter: server::PerUserRateLimiter::new(30, 60),
             oauth_rate_limiter: server::RateLimiter::new(10, 60),
             webhook_rate_limiter: server::RateLimiter::new(10, 60),
             registry_entries: self.state.registry_entries.clone(),
@@ -260,9 +327,15 @@ impl GatewayChannel {
         self
     }
 
-    /// Get the auth token (for printing to console on startup).
+    /// Inject the per-user workspace pool for multi-user mode.
+    pub fn with_workspace_pool(mut self, pool: Arc<server::WorkspacePool>) -> Self {
+        self.rebuild_state(|s| s.workspace_pool = Some(pool));
+        self
+    }
+
+    /// Get the first auth token (for printing to console on startup).
     pub fn auth_token(&self) -> &str {
-        &self.auth_token
+        self.auth.first_token().unwrap_or("")
     }
 
     /// Get a reference to the shared gateway state (for the agent to push SSE events).
@@ -291,7 +364,7 @@ impl Channel for GatewayChannel {
                 ),
             })?;
 
-        server::start_server(addr, self.state.clone(), self.auth_token.clone()).await?;
+        server::start_server(addr, self.state.clone(), self.auth.clone()).await?;
 
         Ok(Box::pin(ReceiverStream::new(rx)))
     }
@@ -311,10 +384,13 @@ impl Channel for GatewayChannel {
             }
         };
 
-        self.state.sse.broadcast(SseEvent::Response {
+        self.state.sse.broadcast_for_user(
-            content: response.content,
+            &msg.user_id,
-            thread_id,
+            AppEvent::Response {
-        });
+                content: response.content,
+                thread_id,
+            },
+        );
 
         Ok(())
     }
@@ -329,11 +405,11 @@ impl Channel for GatewayChannel {
             .and_then(|v| v.as_str())
             .map(String::from);
         let event = match status {
-            StatusUpdate::Thinking(msg) => SseEvent::Thinking {
+            StatusUpdate::Thinking(msg) => AppEvent::Thinking {
                 message: msg,
                 thread_id: thread_id.clone(),
             },
-            StatusUpdate::ToolStarted { name } => SseEvent::ToolStarted {
+            StatusUpdate::ToolStarted { name } => AppEvent::ToolStarted {
                 name,
                 thread_id: thread_id.clone(),
             },
@@ -342,23 +418,23 @@ impl Channel for GatewayChannel {
                 success,
                 error,
                 parameters,
-            } => SseEvent::ToolCompleted {
+            } => AppEvent::ToolCompleted {
                 name,
                 success,
                 error,
                 parameters,
                 thread_id: thread_id.clone(),
             },
-            StatusUpdate::ToolResult { name, preview } => SseEvent::ToolResult {
+            StatusUpdate::ToolResult { name, preview } => AppEvent::ToolResult {
                 name,
                 preview,
                 thread_id: thread_id.clone(),
             },
-            StatusUpdate::StreamChunk(content) => SseEvent::StreamChunk {
+            StatusUpdate::StreamChunk(content) => AppEvent::StreamChunk {
                 content,
                 thread_id: thread_id.clone(),
             },
-            StatusUpdate::Status(msg) => SseEvent::Status {
+            StatusUpdate::Status(msg) => AppEvent::Status {
                 message: msg,
                 thread_id: thread_id.clone(),
             },
@@ -366,7 +442,7 @@ impl Channel for GatewayChannel {
                 job_id,
                 title,
                 browse_url,
-            } => SseEvent::JobStarted {
+            } => AppEvent::JobStarted {
                 job_id,
                 title,
                 browse_url,
@@ -377,7 +453,7 @@ impl Channel for GatewayChannel {
                 description,
                 parameters,
                 allow_always,
-            } => SseEvent::ApprovalNeeded {
+            } => AppEvent::ApprovalNeeded {
                 request_id,
                 tool_name,
                 description,
@@ -391,7 +467,7 @@ impl Channel for GatewayChannel {
                 instructions,
                 auth_url,
                 setup_url,
-            } => SseEvent::AuthRequired {
+            } => AppEvent::AuthRequired {
                 extension_name,
                 instructions,
                 auth_url,
@@ -401,25 +477,39 @@ impl Channel for GatewayChannel {
                 extension_name,
                 success,
                 message,
-            } => SseEvent::AuthCompleted {
+            } => AppEvent::AuthCompleted {
                 extension_name,
                 success,
                 message,
             },
-            StatusUpdate::ImageGenerated { data_url, path } => SseEvent::ImageGenerated {
+            StatusUpdate::ImageGenerated { data_url, path } => AppEvent::ImageGenerated {
                 data_url,
                 path,
                 thread_id: thread_id.clone(),
             },
-            StatusUpdate::Suggestions { suggestions } => SseEvent::Suggestions {
+            StatusUpdate::Suggestions { suggestions } => AppEvent::Suggestions {
                 suggestions,
+                thread_id: thread_id.clone(),
+            },
+            StatusUpdate::ReasoningUpdate {
+                narrative,
+                decisions,
+            } => AppEvent::ReasoningUpdate {
+                narrative,
+                decisions: decisions
+                    .into_iter()
+                    .map(|d| crate::channels::web::types::ToolDecisionDto {
+                        tool_name: d.tool_name,
+                        rationale: d.rationale,
+                    })
+                    .collect(),
                 thread_id,
             },
             StatusUpdate::TurnCost {
                 input_tokens,
                 output_tokens,
                 cost_usd,
-            } => SseEvent::TurnCost {
+            } => AppEvent::TurnCost {
                 input_tokens,
                 output_tokens,
                 cost_usd,
@@ -427,13 +517,21 @@ impl Channel for GatewayChannel {
             },
         };
 
-        self.state.sse.broadcast(event);
+        // Scope events to the user when user_id is available in metadata.
+        // When user_id is missing (heartbeat, routines), events go to all
+        // subscribers. In multi-tenant mode this leaks status across users.
+        if let Some(uid) = metadata.get("user_id").and_then(|v| v.as_str()) {
+            self.state.sse.broadcast_for_user(uid, event);
+        } else {
+            tracing::debug!("Status event missing user_id in metadata; broadcasting globally");
+            self.state.sse.broadcast(event);
+        }
         Ok(())
     }
 
     async fn broadcast(
         &self,
-        _user_id: &str,
+        user_id: &str,
         response: OutgoingResponse,
     ) -> Result<(), ChannelError> {
         let thread_id = match response.thread_id {
@@ -445,10 +543,13 @@ impl Channel for GatewayChannel {
                 return Ok(());
             }
         };
-        self.state.sse.broadcast(SseEvent::Response {
+        self.state.sse.broadcast_for_user(
-            content: response.content,
+            user_id,
-            thread_id,
+            AppEvent::Response {
-        });
+                content: response.content,
+                thread_id,
+            },
+        );
         Ok(())
     }
 

src/channels/web/openai_compat.rs

@@ -231,6 +231,7 @@ pub fn convert_messages(messages: &[OpenAiMessage]) -> Result<Vec<ChatMessage>,
                                 name: tc.function.name.clone(),
                                 arguments: serde_json::from_str(&tc.function.arguments)
                                     .unwrap_or(serde_json::Value::Object(Default::default())),
+                                reasoning: None,
                             })
                             .collect();
                         Ok(ChatMessage::assistant_with_tool_calls(
@@ -463,9 +464,10 @@ fn build_tool_request(
 
 pub async fn chat_completions_handler(
     State(state): State<Arc<GatewayState>>,
+    super::auth::AuthenticatedUser(user): super::auth::AuthenticatedUser,
     Json(req): Json<OpenAiChatRequest>,
 ) -> Result<impl IntoResponse, (StatusCode, Json<OpenAiErrorResponse>)> {
-    if !state.chat_rate_limiter.check() {
+    if !state.chat_rate_limiter.check(&user.user_id) {
         return Err(openai_error(
             StatusCode::TOO_MANY_REQUESTS,
             "Rate limit exceeded. Please try again later.",
@@ -953,6 +955,7 @@ mod tests {
             id: "call_abc".to_string(),
             name: "search".to_string(),
             arguments: serde_json::json!({"query": "rust"}),
+            reasoning: None,
         }];
 
         let converted = convert_tool_calls_to_openai(&calls);

src/channels/web/sse.rs

@@ -11,15 +11,31 @@ use tokio::sync::broadcast;
 use tokio_stream::StreamExt;
 use tokio_stream::wrappers::BroadcastStream;
 
-use crate::channels::web::types::SseEvent;
+use crate::channels::web::types::AppEvent;
 
 /// Maximum number of concurrent SSE/WebSocket connections.
 /// Prevents resource exhaustion from connection flooding.
 const MAX_CONNECTIONS: u64 = 100;
 
+/// Envelope for broadcast events: carries an optional user scope.
+///
+/// `user_id = None` means the event is global (e.g. Heartbeat) and delivered
+/// to all subscribers. `user_id = Some(id)` means the event is only delivered
+/// to subscribers that match that user_id.
+#[derive(Debug, Clone)]
+pub(crate) struct ScopedEvent {
+    pub(crate) user_id: Option<String>,
+    pub(crate) event: AppEvent,
+}
+
 /// Manages SSE broadcast to all connected browser tabs.
+///
+/// In multi-user mode, events are scoped by user_id so that each subscriber
+/// only receives events intended for their user (plus global events like
+/// Heartbeat). In single-user mode, all events are delivered to all subscribers
+/// (backwards compatible).
 pub struct SseManager {
-    tx: broadcast::Sender<SseEvent>,
+    tx: broadcast::Sender<ScopedEvent>,
     connection_count: Arc<AtomicU64>,
     max_connections: u64,
 }
@@ -45,7 +61,7 @@ impl SseManager {
     /// only be called before the server starts accepting connections (i.e.,
     /// during startup wiring). Calling it after connections are established
     /// will break connection tracking and allow exceeding `MAX_CONNECTIONS`.
-    pub fn from_sender(tx: broadcast::Sender<SseEvent>) -> Self {
+    pub(crate) fn from_sender(tx: broadcast::Sender<ScopedEvent>) -> Self {
         Self {
             tx,
             connection_count: Arc::new(AtomicU64::new(0)),
@@ -53,15 +69,28 @@ impl SseManager {
         }
     }
 
-    /// Broadcast an event to all connected clients.
+    /// Get a clone of the broadcast sender for use by other components.
-    pub fn broadcast(&self, event: SseEvent) {
+    pub(crate) fn sender(&self) -> broadcast::Sender<ScopedEvent> {
-        // Ignore send errors (no receivers is fine)
+        self.tx.clone()
-        let _ = self.tx.send(event);
     }
 
-    /// Get a clone of the broadcast sender for use by other components.
+    /// Broadcast an event to all connected clients (global/unscoped).
-    pub fn sender(&self) -> broadcast::Sender<SseEvent> {
+    pub fn broadcast(&self, event: AppEvent) {
-        self.tx.clone()
+        let _ = self.tx.send(ScopedEvent {
+            user_id: None,
+            event,
+        });
+    }
+
+    /// Broadcast an event scoped to a specific user.
+    ///
+    /// Only subscribers for this user_id (or unscoped subscribers) will
+    /// receive the event.
+    pub fn broadcast_for_user(&self, user_id: &str, event: AppEvent) {
+        let _ = self.tx.send(ScopedEvent {
+            user_id: Some(user_id.to_string()),
+            event,
+        });
     }
 
     /// Get current number of active connections.
@@ -71,11 +100,15 @@ impl SseManager {
 
     /// Create a raw broadcast subscription for non-SSE consumers (e.g. WebSocket).
     ///
-    /// Returns a stream of `SseEvent` values and increments/decrements the
+    /// When `user_id` is `Some`, only events scoped to that user (or global
-    /// connection counter on creation/drop, just like `subscribe()` does for SSE.
+    /// events) are delivered. When `None`, all events are delivered (single-user
+    /// backwards compatibility).
     ///
     /// Returns `None` if the maximum connection limit has been reached.
-    pub fn subscribe_raw(&self) -> Option<impl Stream<Item = SseEvent> + Send + 'static + use<>> {
+    pub fn subscribe_raw(
+        &self,
+        user_id: Option<String>,
+    ) -> Option<impl Stream<Item = AppEvent> + Send + 'static + use<>> {
         // Atomically increment only if below the limit. This prevents
         // concurrent callers from overshooting max_connections.
         let counter = Arc::clone(&self.connection_count);
@@ -91,7 +124,19 @@ impl SseManager {
             .ok()?;
         let rx = self.tx.subscribe();
 
-        let stream = BroadcastStream::new(rx).filter_map(|result| result.ok());
+        let stream = BroadcastStream::new(rx).filter_map(move |result| match result {
+            Ok(scoped) => {
+                // Global events (user_id=None) always pass through.
+                // Scoped events only pass if the subscriber matches (or subscriber is unscoped).
+                match (&user_id, &scoped.user_id) {
+                    (_, None) => Some(scoped.event), // global -> all
+                    (None, _) => Some(scoped.event), // unscoped subscriber -> all
+                    (Some(sub), Some(ev)) if sub == ev => Some(scoped.event), // match
+                    _ => None,                       // different user -> skip
+                }
+            }
+            Err(_) => None,
+        });
 
         Some(CountedStream {
             inner: stream,
@@ -101,9 +146,13 @@ impl SseManager {
 
     /// Create a new SSE stream for a client connection.
     ///
+    /// When `user_id` is `Some`, only events for that user (or global events)
+    /// are delivered. When `None`, all events are delivered.
+    ///
     /// Returns `None` if the maximum connection limit has been reached.
     pub fn subscribe(
         &self,
+        user_id: Option<String>,
     ) -> Option<Sse<impl Stream<Item = Result<Event, Infallible>> + Send + 'static + use<>>> {
         // Atomically increment only if below the limit.
         let counter = Arc::clone(&self.connection_count);
@@ -120,34 +169,25 @@ impl SseManager {
         let rx = self.tx.subscribe();
 
         let stream = BroadcastStream::new(rx)
-            .filter_map(|result| result.ok())
+            .filter_map(move |result| match result {
-            .map(|event| {
+                Ok(scoped) => match (&user_id, &scoped.user_id) {
-                let data = serde_json::to_string(&event).unwrap_or_default();
+                    (_, None) => Some(scoped.event),
-                let event_type = match &event {
+                    (None, _) => Some(scoped.event),
-                    SseEvent::Response { .. } => "response",
+                    (Some(sub), Some(ev)) if sub == ev => Some(scoped.event),
-                    SseEvent::Thinking { .. } => "thinking",
+                    _ => None,
-                    SseEvent::ToolStarted { .. } => "tool_started",
+                },
-                    SseEvent::ToolCompleted { .. } => "tool_completed",
+                Err(_) => None,
-                    SseEvent::ToolResult { .. } => "tool_result",
+            })
-                    SseEvent::StreamChunk { .. } => "stream_chunk",
+            .filter_map(|event| {
-                    SseEvent::Status { .. } => "status",
+                let data = match serde_json::to_string(&event) {
-                    SseEvent::ApprovalNeeded { .. } => "approval_needed",
+                    Ok(s) => s,
-                    SseEvent::AuthRequired { .. } => "auth_required",
+                    Err(e) => {
-                    SseEvent::AuthCompleted { .. } => "auth_completed",
+                        tracing::warn!("Failed to serialize SSE event: {}", e);
-                    SseEvent::Error { .. } => "error",
+                        return None;
-                    SseEvent::JobStarted { .. } => "job_started",
+                    }
-                    SseEvent::JobMessage { .. } => "job_message",
-                    SseEvent::JobToolUse { .. } => "job_tool_use",
-                    SseEvent::JobToolResult { .. } => "job_tool_result",
-                    SseEvent::JobStatus { .. } => "job_status",
-                    SseEvent::JobResult { .. } => "job_result",
-                    SseEvent::Heartbeat => "heartbeat",
-                    SseEvent::ImageGenerated { .. } => "image_generated",
-                    SseEvent::Suggestions { .. } => "suggestions",
-                    SseEvent::TurnCost { .. } => "turn_cost",
-                    SseEvent::ExtensionStatus { .. } => "extension_status",
                 };
-                Ok(Event::default().event(event_type).data(data))
+                let event_type = event.event_type();
+                Some(Ok(Event::default().event(event_type).data(data)))
             });
 
         // Wrap in a stream that decrements on drop
@@ -209,24 +249,22 @@ mod tests {
     fn test_broadcast_without_receivers() {
         let manager = SseManager::new();
         // Should not panic even with no receivers
-        manager.broadcast(SseEvent::Heartbeat);
+        manager.broadcast(AppEvent::Heartbeat);
     }
 
     #[tokio::test]
     async fn test_broadcast_to_receiver() {
         let manager = SseManager::new();
-        let mut rx = BroadcastStream::new(manager.tx.subscribe());
+        let mut stream = Box::pin(manager.subscribe_raw(None).expect("should subscribe"));
 
-        manager.broadcast(SseEvent::Status {
+        manager.broadcast(AppEvent::Status {
             message: "test".to_string(),
             thread_id: None,
         });
 
-        let event = rx.next().await;
+        let event = stream.next().await.unwrap();
-        assert!(event.is_some());
-        let event = event.unwrap().unwrap();
         match event {
-            SseEvent::Status { message, .. } => assert_eq!(message, "test"),
+            AppEvent::Status { message, .. } => assert_eq!(message, "test"),
             _ => panic!("unexpected event type"),
         }
     }
@@ -234,18 +272,18 @@ mod tests {
     #[tokio::test]
     async fn test_subscribe_raw_receives_events() {
         let manager = SseManager::new();
-        let mut stream = Box::pin(manager.subscribe_raw().expect("should subscribe"));
+        let mut stream = Box::pin(manager.subscribe_raw(None).expect("should subscribe"));
 
         assert_eq!(manager.connection_count(), 1);
 
-        manager.broadcast(SseEvent::Thinking {
+        manager.broadcast(AppEvent::Thinking {
             message: "working".to_string(),
             thread_id: None,
         });
 
         let event = stream.next().await.unwrap();
         match event {
-            SseEvent::Thinking { message, .. } => assert_eq!(message, "working"),
+            AppEvent::Thinking { message, .. } => assert_eq!(message, "working"),
             _ => panic!("Expected Thinking event"),
         }
     }
@@ -254,7 +292,7 @@ mod tests {
     async fn test_subscribe_raw_decrements_on_drop() {
         let manager = SseManager::new();
         {
-            let _stream = Box::pin(manager.subscribe_raw().expect("should subscribe"));
+            let _stream = Box::pin(manager.subscribe_raw(None).expect("should subscribe"));
             assert_eq!(manager.connection_count(), 1);
         }
         // Stream dropped, counter should decrement
@@ -264,16 +302,16 @@ mod tests {
     #[tokio::test]
     async fn test_subscribe_raw_multiple_subscribers() {
         let manager = SseManager::new();
-        let mut s1 = Box::pin(manager.subscribe_raw().expect("should subscribe"));
+        let mut s1 = Box::pin(manager.subscribe_raw(None).expect("should subscribe"));
-        let mut s2 = Box::pin(manager.subscribe_raw().expect("should subscribe"));
+        let mut s2 = Box::pin(manager.subscribe_raw(None).expect("should subscribe"));
         assert_eq!(manager.connection_count(), 2);
 
-        manager.broadcast(SseEvent::Heartbeat);
+        manager.broadcast(AppEvent::Heartbeat);
 
         let e1 = s1.next().await.unwrap();
         let e2 = s2.next().await.unwrap();
-        assert!(matches!(e1, SseEvent::Heartbeat));
+        assert!(matches!(e1, AppEvent::Heartbeat));
-        assert!(matches!(e2, SseEvent::Heartbeat));
+        assert!(matches!(e2, AppEvent::Heartbeat));
 
         drop(s1);
         assert_eq!(manager.connection_count(), 1);
@@ -286,12 +324,51 @@ mod tests {
         let mut manager = SseManager::new();
         manager.max_connections = 2; // Low limit for testing
 
-        let _s1 = Box::pin(manager.subscribe_raw().expect("first should succeed"));
+        let _s1 = Box::pin(manager.subscribe_raw(None).expect("first should succeed"));
-        let _s2 = Box::pin(manager.subscribe_raw().expect("second should succeed"));
+        let _s2 = Box::pin(manager.subscribe_raw(None).expect("second should succeed"));
         assert_eq!(manager.connection_count(), 2);
 
         // Third should be rejected
-        assert!(manager.subscribe_raw().is_none());
+        assert!(manager.subscribe_raw(None).is_none());
-        assert!(manager.subscribe().is_none());
+        assert!(manager.subscribe(None).is_none());
+    }
+
+    #[tokio::test]
+    async fn test_scoped_events_filtered_by_user() {
+        let manager = SseManager::new();
+        let mut alice = Box::pin(
+            manager
+                .subscribe_raw(Some("alice".to_string()))
+                .expect("subscribe"),
+        );
+        let mut bob = Box::pin(
+            manager
+                .subscribe_raw(Some("bob".to_string()))
+                .expect("subscribe"),
+        );
+
+        // Send event scoped to alice
+        manager.broadcast_for_user(
+            "alice",
+            AppEvent::Status {
+                message: "alice only".to_string(),
+                thread_id: None,
+            },
+        );
+
+        // Send global event
+        manager.broadcast(AppEvent::Heartbeat);
+
+        // Alice gets her scoped event
+        let e = alice.next().await.unwrap();
+        assert!(matches!(e, AppEvent::Status { .. }));
+
+        // Alice also gets the global heartbeat
+        let e = alice.next().await.unwrap();
+        assert!(matches!(e, AppEvent::Heartbeat));
+
+        // Bob only gets the global heartbeat (alice's event was filtered)
+        let e = bob.next().await.unwrap(); // safety: test-only
+        assert!(matches!(e, AppEvent::Heartbeat)); // safety: test assertion
     }
 }

src/channels/web/static/app.js

@@ -4265,9 +4265,9 @@ function renderRoutineDetail(routine) {
       + '<th>Trigger</th><th>Started</th><th>Completed</th><th>Status</th><th>Summary</th><th>Tokens</th>'
       + '</tr></thead><tbody>';
     for (const run of routine.recent_runs) {
-      const runStatusClass = run.status === 'Ok' ? 'completed'
+      const runStatusClass = run.status === 'ok' ? 'completed'
-        : run.status === 'Failed' ? 'failed'
+        : run.status === 'failed' ? 'failed'
-        : run.status === 'Attention' ? 'stuck'
+        : run.status === 'attention' ? 'stuck'
         : 'in_progress';
       html += '<tr>'
         + '<td>' + escapeHtml(run.trigger_type) + '</td>'

src/channels/web/test_helpers.rs

@@ -10,7 +10,8 @@ use std::sync::Arc;
 use tokio::sync::mpsc;
 
 use crate::channels::IncomingMessage;
-use crate::channels::web::server::{GatewayState, RateLimiter, start_server};
+use crate::channels::web::auth::MultiAuthState;
+use crate::channels::web::server::{GatewayState, PerUserRateLimiter, RateLimiter, start_server};
 use crate::channels::web::sse::SseManager;
 use crate::channels::web::ws::WsConnectionTracker;
 
@@ -64,8 +65,9 @@ impl TestGatewayBuilder {
     pub fn build(self) -> Arc<GatewayState> {
         Arc::new(GatewayState {
             msg_tx: tokio::sync::RwLock::new(self.msg_tx),
-            sse: SseManager::new(),
+            sse: Arc::new(SseManager::new()),
             workspace: None,
+            workspace_pool: None,
             session_manager: None,
             log_broadcaster: None,
             log_level_handle: None,
@@ -74,14 +76,15 @@ impl TestGatewayBuilder {
             store: None,
             job_manager: None,
             prompt_queue: None,
-            user_id: self.user_id,
+            owner_id: self.user_id.clone(),
+            default_sender_id: self.user_id,
             shutdown_tx: tokio::sync::RwLock::new(None),
             ws_tracker: Some(Arc::new(WsConnectionTracker::new())),
             llm_provider: self.llm_provider,
             skill_registry: None,
             skill_catalog: None,
             scheduler: None,
-            chat_rate_limiter: RateLimiter::new(30, 60),
+            chat_rate_limiter: PerUserRateLimiter::new(30, 60),
             oauth_rate_limiter: RateLimiter::new(10, 60),
             webhook_rate_limiter: RateLimiter::new(10, 60),
             registry_entries: Vec::new(),
@@ -98,11 +101,26 @@ impl TestGatewayBuilder {
         self,
         auth_token: &str,
     ) -> Result<(SocketAddr, Arc<GatewayState>), crate::error::ChannelError> {
+        let auth = MultiAuthState::single(auth_token.to_string(), "test-user".to_string());
         let state = self.build();
         let addr: SocketAddr = "127.0.0.1:0"
             .parse()
-            .expect("hard-coded address must parse");
+            .expect("hard-coded address must parse"); // safety: constant literal
-        let bound = start_server(addr, state.clone(), auth_token.to_string()).await?;
+        let bound = start_server(addr, state.clone(), auth).await?;
+        Ok((bound, state))
+    }
+
+    /// Build the state and start a gateway server with multi-user auth.
+    /// Returns the bound address and the shared state.
+    pub async fn start_multi(
+        self,
+        auth: MultiAuthState,
+    ) -> Result<(SocketAddr, Arc<GatewayState>), crate::error::ChannelError> {
+        let state = self.build();
+        let addr: SocketAddr = "127.0.0.1:0"
+            .parse()
+            .expect("hard-coded address must parse"); // safety: constant literal
+        let bound = start_server(addr, state.clone(), auth).await?;
         Ok((bound, state))
     }
 }

src/channels/web/tests/mod.rs

@@ -0,0 +1,3 @@
+//! Integration tests for the web gateway module.
+
+mod multi_tenant;

src/channels/web/tests/multi_tenant.rs

@@ -0,0 +1,833 @@
+//! Multi-tenant isolation tests for the web gateway.
+//!
+//! Tests cover workspace pool scoping, job handler isolation, and auth
+//! enforcement on protected endpoints. Uses `LibSqlBackend::new_local()`
+//! with a temporary directory for a real (but ephemeral) database.
+
+use std::collections::HashMap;
+use std::sync::Arc;
+use std::time::Duration;
+
+use axum::Router;
+use axum::body::Body;
+use axum::http::{Method, Request, StatusCode};
+use axum::middleware;
+use axum::routing::{delete, get, post};
+use tower::ServiceExt;
+use uuid::Uuid;
+
+use crate::channels::web::GatewayChannel;
+use crate::channels::web::auth::{
+    AuthenticatedUser, MultiAuthState, UserIdentity, auth_middleware,
+};
+use crate::channels::web::server::{
+    ActiveConfigSnapshot, GatewayState, PerUserRateLimiter, PromptQueue, RateLimiter, WorkspacePool,
+};
+use crate::channels::web::sse::SseManager;
+use crate::config::GatewayConfig;
+
+// ── Helpers ────────────────────────────────────────────────────────────
+
+/// Create a two-user `MultiAuthState` for alice and bob.
+fn two_user_auth() -> MultiAuthState {
+    let mut tokens = HashMap::new();
+    tokens.insert(
+        "tok-alice".to_string(),
+        UserIdentity {
+            user_id: "alice".to_string(),
+            workspace_read_scopes: vec!["shared".to_string()],
+        },
+    );
+    tokens.insert(
+        "tok-bob".to_string(),
+        UserIdentity {
+            user_id: "bob".to_string(),
+            workspace_read_scopes: vec!["shared".to_string(), "alice".to_string()],
+        },
+    );
+    MultiAuthState::multi(tokens)
+}
+
+/// Build a `GatewayState` with configurable store and prompt queue.
+fn build_state(
+    store: Option<Arc<dyn crate::db::Database>>,
+    prompt_queue: Option<PromptQueue>,
+) -> Arc<GatewayState> {
+    Arc::new(GatewayState {
+        msg_tx: tokio::sync::RwLock::new(None),
+        sse: Arc::new(SseManager::new()),
+        workspace: None,
+        workspace_pool: None,
+        session_manager: None,
+        log_broadcaster: None,
+        log_level_handle: None,
+        extension_manager: None,
+        tool_registry: None,
+        store,
+        job_manager: None,
+        prompt_queue,
+        owner_id: "test".to_string(),
+        default_sender_id: "test".to_string(),
+        shutdown_tx: tokio::sync::RwLock::new(None),
+        ws_tracker: None,
+        llm_provider: None,
+        skill_registry: None,
+        skill_catalog: None,
+        scheduler: None,
+        chat_rate_limiter: PerUserRateLimiter::new(30, 60),
+        oauth_rate_limiter: RateLimiter::new(10, 60),
+        webhook_rate_limiter: RateLimiter::new(10, 60),
+        registry_entries: Vec::new(),
+        cost_guard: None,
+        routine_engine: Arc::new(tokio::sync::RwLock::new(None)),
+        startup_time: std::time::Instant::now(),
+        active_config: ActiveConfigSnapshot::default(),
+    })
+}
+
+fn gateway_config() -> GatewayConfig {
+    GatewayConfig {
+        host: "127.0.0.1".to_string(),
+        port: 3000,
+        auth_token: Some("gateway-auth".to_string()),
+        user_id: "gateway-sender".to_string(),
+        workspace_read_scopes: Vec::new(),
+        memory_layers: Vec::new(),
+        user_tokens: None,
+    }
+}
+
+#[test]
+fn with_owner_scope_updates_gateway_owner_scope_in_multi_user_mode() {
+    let mut gateway = GatewayChannel::new(gateway_config());
+    gateway.auth = two_user_auth();
+    gateway.config.user_tokens = Some(HashMap::new());
+    let gateway = gateway.with_owner_scope("owner-scope");
+
+    assert_eq!(gateway.state.owner_id, "owner-scope");
+    assert_eq!(gateway.state.default_sender_id, "gateway-sender");
+
+    let alice = gateway
+        .auth
+        .authenticate("tok-alice")
+        .expect("alice token should remain valid");
+    let bob = gateway
+        .auth
+        .authenticate("tok-bob")
+        .expect("bob token should remain valid");
+    assert_eq!(alice.user_id, "alice");
+    assert_eq!(bob.user_id, "bob");
+}
+
+/// Create a libSQL-backed test database in a temporary directory.
+///
+/// Returns the database and a `TempDir` guard — the database file is
+/// deleted when the guard is dropped.
+#[cfg(feature = "libsql")]
+async fn test_db() -> (Arc<dyn crate::db::Database>, tempfile::TempDir) {
+    use crate::db::Database;
+    let dir = tempfile::tempdir().expect("failed to create temp dir"); // safety: test-only
+    let path = dir.path().join("test.db");
+    let backend = crate::db::libsql::LibSqlBackend::new_local(&path)
+        .await
+        .expect("failed to create test LibSqlBackend"); // safety: test-only
+    backend
+        .run_migrations()
+        .await
+        .expect("failed to run migrations"); // safety: test-only
+    (Arc::new(backend) as Arc<dyn crate::db::Database>, dir)
+}
+
+/// Build a minimal Routine for testing.
+fn make_routine(user_id: &str, name: &str) -> crate::agent::routine::Routine {
+    let now = chrono::Utc::now();
+    crate::agent::routine::Routine {
+        id: Uuid::new_v4(),
+        name: name.to_string(),
+        description: format!("Test routine: {name}"),
+        user_id: user_id.to_string(),
+        enabled: true,
+        trigger: crate::agent::routine::Trigger::Cron {
+            schedule: "0 9 * * *".to_string(),
+            timezone: None,
+        },
+        action: crate::agent::routine::RoutineAction::Lightweight {
+            prompt: "hello".to_string(),
+            context_paths: vec![],
+            max_tokens: 1024,
+            use_tools: false,
+            max_tool_rounds: 3,
+        },
+        guardrails: crate::agent::routine::RoutineGuardrails {
+            cooldown: Duration::from_secs(60),
+            max_concurrent: 1,
+            dedup_window: None,
+        },
+        notify: crate::agent::routine::NotifyConfig {
+            channel: None,
+            user: None,
+            on_success: false,
+            on_failure: true,
+            on_attention: true,
+        },
+        last_run_at: None,
+        next_fire_at: None,
+        run_count: 0,
+        consecutive_failures: 0,
+        state: serde_json::json!({}),
+        created_at: now,
+        updated_at: now,
+    }
+}
+
+/// Build a minimal SandboxJobRecord for testing.
+fn make_sandbox_job(user_id: &str, task: &str) -> crate::history::SandboxJobRecord {
+    let now = chrono::Utc::now();
+    crate::history::SandboxJobRecord {
+        id: Uuid::new_v4(),
+        task: task.to_string(),
+        status: "completed".to_string(),
+        user_id: user_id.to_string(),
+        project_dir: format!("/tmp/test-{}", Uuid::new_v4()),
+        success: Some(true),
+        failure_reason: None,
+        created_at: now,
+        started_at: Some(now),
+        completed_at: Some(now),
+        credential_grants_json: "[]".to_string(),
+    }
+}
+
+// ═══════════════════════════════════════════════════════════════════════
+// WorkspacePool Tests
+// ═══════════════════════════════════════════════════════════════════════
+
+#[cfg(feature = "libsql")]
+mod workspace_pool {
+    use super::*;
+    use crate::config::{WorkspaceConfig, WorkspaceSearchConfig};
+    use crate::workspace::EmbeddingCacheConfig;
+    use crate::workspace::layer::MemoryLayer;
+
+    #[tokio::test]
+    async fn test_workspace_pool_applies_search_config() {
+        let (db, _dir) = test_db().await;
+        let search_config = WorkspaceSearchConfig {
+            rrf_k: 42,
+            ..Default::default()
+        };
+        let pool = WorkspacePool::new(
+            db,
+            None,
+            EmbeddingCacheConfig::default(),
+            search_config,
+            WorkspaceConfig::default(),
+        );
+        let identity = UserIdentity {
+            user_id: "alice".to_string(),
+            workspace_read_scopes: vec![],
+        };
+        let ws = pool.get_or_create(&identity).await;
+        assert_eq!(ws.user_id(), "alice");
+    }
+
+    #[tokio::test]
+    async fn test_workspace_pool_applies_memory_layers() {
+        let (db, _dir) = test_db().await;
+        let layers = vec![MemoryLayer {
+            name: "shared-layer".to_string(),
+            scope: "shared".to_string(),
+            writable: false,
+            sensitivity: Default::default(),
+        }];
+        let ws_config = WorkspaceConfig {
+            memory_layers: layers,
+            read_scopes: vec![],
+        };
+        let pool = WorkspacePool::new(
+            db,
+            None,
+            EmbeddingCacheConfig::default(),
+            WorkspaceSearchConfig::default(),
+            ws_config,
+        );
+        let identity = UserIdentity {
+            user_id: "alice".to_string(),
+            workspace_read_scopes: vec![],
+        };
+        let ws = pool.get_or_create(&identity).await;
+        // Memory layer scope "shared" should appear in read_user_ids.
+        assert!(
+            ws.read_user_ids().contains(&"shared".to_string()),
+            "expected 'shared' in read_user_ids, got {:?}",
+            ws.read_user_ids()
+        );
+    }
+
+    #[tokio::test]
+    async fn test_workspace_pool_applies_identity_read_scopes() {
+        let (db, _dir) = test_db().await;
+        let pool = WorkspacePool::new(
+            db,
+            None,
+            EmbeddingCacheConfig::default(),
+            WorkspaceSearchConfig::default(),
+            WorkspaceConfig::default(),
+        );
+        let identity = UserIdentity {
+            user_id: "bob".to_string(),
+            workspace_read_scopes: vec!["alice".to_string(), "shared".to_string()],
+        };
+        let ws = pool.get_or_create(&identity).await;
+        assert_eq!(ws.user_id(), "bob");
+        assert!(
+            ws.read_user_ids().contains(&"alice".to_string()),
+            "expected 'alice' in read_user_ids from identity scopes"
+        );
+        assert!(
+            ws.read_user_ids().contains(&"shared".to_string()),
+            "expected 'shared' in read_user_ids from identity scopes"
+        );
+    }
+
+    #[tokio::test]
+    async fn test_workspace_pool_caches_per_user() {
+        let (db, _dir) = test_db().await;
+        let pool = WorkspacePool::new(
+            db,
+            None,
+            EmbeddingCacheConfig::default(),
+            WorkspaceSearchConfig::default(),
+            WorkspaceConfig::default(),
+        );
+        let alice_id = UserIdentity {
+            user_id: "alice".to_string(),
+            workspace_read_scopes: vec![],
+        };
+        let bob_id = UserIdentity {
+            user_id: "bob".to_string(),
+            workspace_read_scopes: vec![],
+        };
+
+        let alice_ws1 = pool.get_or_create(&alice_id).await;
+        let alice_ws2 = pool.get_or_create(&alice_id).await;
+        let bob_ws = pool.get_or_create(&bob_id).await;
+
+        // Same user gets the same Arc.
+        assert!(Arc::ptr_eq(&alice_ws1, &alice_ws2));
+        // Different users get different instances.
+        assert!(!Arc::ptr_eq(&alice_ws1, &bob_ws));
+        assert_eq!(alice_ws1.user_id(), "alice");
+        assert_eq!(bob_ws.user_id(), "bob");
+    }
+
+    #[tokio::test]
+    async fn test_workspace_pool_combines_global_and_identity_scopes() {
+        let (db, _dir) = test_db().await;
+        let ws_config = WorkspaceConfig {
+            memory_layers: vec![],
+            read_scopes: vec!["global-shared".to_string()],
+        };
+        let pool = WorkspacePool::new(
+            db,
+            None,
+            EmbeddingCacheConfig::default(),
+            WorkspaceSearchConfig::default(),
+            ws_config,
+        );
+        let identity = UserIdentity {
+            user_id: "alice".to_string(),
+            workspace_read_scopes: vec!["token-scope".to_string()],
+        };
+        let ws = pool.get_or_create(&identity).await;
+        let scopes = ws.read_user_ids();
+        // Primary scope
+        assert!(scopes.contains(&"alice".to_string()));
+        // Global config scope
+        assert!(
+            scopes.contains(&"global-shared".to_string()),
+            "expected global scope 'global-shared', got {:?}",
+            scopes
+        );
+        // Token identity scope
+        assert!(
+            scopes.contains(&"token-scope".to_string()),
+            "expected token scope 'token-scope', got {:?}",
+            scopes
+        );
+    }
+}
+
+// ═══════════════════════════════════════════════════════════════════════
+// Jobs Handler Isolation Tests
+// ═══════════════════════════════════════════════════════════════════════
+
+#[cfg(feature = "libsql")]
+mod jobs_isolation {
+    use super::*;
+    use crate::channels::web::handlers::jobs::{
+        jobs_cancel_handler, jobs_prompt_handler, jobs_restart_handler, jobs_summary_handler,
+    };
+    // SandboxStore methods are accessed through the Database supertrait.
+
+    /// Build a router with job endpoints behind multi-user auth.
+    fn jobs_router(state: Arc<GatewayState>, auth: MultiAuthState) -> Router {
+        Router::new()
+            .route("/api/jobs/summary", get(jobs_summary_handler))
+            .route("/api/jobs/{id}/cancel", post(jobs_cancel_handler))
+            .route("/api/jobs/{id}/restart", post(jobs_restart_handler))
+            .route("/api/jobs/{id}/prompt", post(jobs_prompt_handler))
+            .layer(middleware::from_fn_with_state(auth, auth_middleware))
+            .with_state(state)
+    }
+
+    #[tokio::test]
+    async fn test_jobs_summary_scoped_to_user() {
+        let (db, _dir) = test_db().await;
+
+        // Insert sandbox jobs for alice and bob.
+        let alice_job = make_sandbox_job("alice", "alice task");
+        let bob_job = make_sandbox_job("bob", "bob task");
+        db.save_sandbox_job(&alice_job).await.unwrap();
+        db.save_sandbox_job(&bob_job).await.unwrap();
+
+        let state = build_state(Some(db), None);
+        let auth = two_user_auth();
+        let app = jobs_router(state, auth);
+
+        // Alice should see 1 job.
+        let req = Request::builder()
+            .uri("/api/jobs/summary")
+            .header("Authorization", "Bearer tok-alice")
+            .body(Body::empty())
+            .unwrap();
+        let resp = app.clone().oneshot(req).await.unwrap();
+        assert_eq!(resp.status(), StatusCode::OK);
+        let body: serde_json::Value =
+            serde_json::from_slice(&axum::body::to_bytes(resp.into_body(), 4096).await.unwrap())
+                .unwrap();
+        assert_eq!(body["total"], 1, "alice should see only her own jobs");
+
+        // Bob should see 1 job.
+        let req = Request::builder()
+            .uri("/api/jobs/summary")
+            .header("Authorization", "Bearer tok-bob")
+            .body(Body::empty())
+            .unwrap();
+        let resp = app.oneshot(req).await.unwrap();
+        assert_eq!(resp.status(), StatusCode::OK);
+        let body: serde_json::Value =
+            serde_json::from_slice(&axum::body::to_bytes(resp.into_body(), 4096).await.unwrap())
+                .unwrap();
+        assert_eq!(body["total"], 1, "bob should see only his own jobs");
+    }
+
+    #[tokio::test]
+    async fn test_jobs_restart_rejects_other_user() {
+        let (db, _dir) = test_db().await;
+
+        // Insert a failed sandbox job owned by alice.
+        let mut alice_job = make_sandbox_job("alice", "alice task");
+        alice_job.status = "failed".to_string();
+        alice_job.success = Some(false);
+        db.save_sandbox_job(&alice_job).await.unwrap();
+
+        let state = build_state(Some(db), None);
+        let auth = two_user_auth();
+        let app = jobs_router(state, auth);
+
+        // Bob tries to restart alice's job.
+        let req = Request::builder()
+            .method(Method::POST)
+            .uri(format!("/api/jobs/{}/restart", alice_job.id))
+            .header("Authorization", "Bearer tok-bob")
+            .body(Body::empty())
+            .unwrap();
+        let resp = app.oneshot(req).await.unwrap();
+        assert_eq!(
+            resp.status(),
+            StatusCode::NOT_FOUND,
+            "bob should not be able to restart alice's job"
+        );
+    }
+
+    #[tokio::test]
+    async fn test_jobs_prompt_works_for_agent_jobs() {
+        let (db, _dir) = test_db().await;
+
+        // Insert a running sandbox job owned by alice in claude_code mode.
+        let mut alice_job = make_sandbox_job("alice", "prompt test");
+        alice_job.status = "running".to_string();
+        alice_job.success = None;
+        alice_job.completed_at = None;
+        db.save_sandbox_job(&alice_job).await.unwrap();
+        db.update_sandbox_job_mode(alice_job.id, "claude_code")
+            .await
+            .unwrap();
+
+        let prompt_queue: PromptQueue =
+            Arc::new(tokio::sync::Mutex::new(std::collections::HashMap::new()));
+        let state = build_state(Some(db), Some(prompt_queue.clone()));
+        let auth = two_user_auth();
+        let app = jobs_router(state, auth);
+
+        // Alice prompts her own job.
+        let req = Request::builder()
+            .method(Method::POST)
+            .uri(format!("/api/jobs/{}/prompt", alice_job.id))
+            .header("Authorization", "Bearer tok-alice")
+            .header("Content-Type", "application/json")
+            .body(Body::from(
+                serde_json::to_string(&serde_json::json!({"content": "hello"})).unwrap(),
+            ))
+            .unwrap();
+        let resp = app.oneshot(req).await.unwrap();
+        assert_eq!(
+            resp.status(),
+            StatusCode::OK,
+            "alice should be able to prompt her own job"
+        );
+
+        // Verify prompt was enqueued.
+        let queue = prompt_queue.lock().await;
+        assert!(
+            queue.contains_key(&alice_job.id),
+            "prompt queue should contain alice's job"
+        );
+    }
+
+    #[tokio::test]
+    async fn test_jobs_prompt_rejects_other_user() {
+        let (db, _dir) = test_db().await;
+
+        let mut alice_job = make_sandbox_job("alice", "alice task");
+        alice_job.status = "running".to_string();
+        alice_job.success = None;
+        alice_job.completed_at = None;
+        db.save_sandbox_job(&alice_job).await.unwrap();
+        db.update_sandbox_job_mode(alice_job.id, "claude_code")
+            .await
+            .unwrap();
+
+        let prompt_queue: PromptQueue =
+            Arc::new(tokio::sync::Mutex::new(std::collections::HashMap::new()));
+        let state = build_state(Some(db), Some(prompt_queue));
+        let auth = two_user_auth();
+        let app = jobs_router(state, auth);
+
+        // Bob tries to prompt alice's job.
+        let req = Request::builder()
+            .method(Method::POST)
+            .uri(format!("/api/jobs/{}/prompt", alice_job.id))
+            .header("Authorization", "Bearer tok-bob")
+            .header("Content-Type", "application/json")
+            .body(Body::from(
+                serde_json::to_string(&serde_json::json!({"content": "sneaky"})).unwrap(),
+            ))
+            .unwrap();
+        let resp = app.oneshot(req).await.unwrap();
+        assert_eq!(
+            resp.status(),
+            StatusCode::NOT_FOUND,
+            "bob should not be able to prompt alice's job"
+        );
+    }
+
+    #[tokio::test]
+    async fn test_jobs_cancel_rejects_other_user() {
+        let (db, _dir) = test_db().await;
+
+        let mut alice_job = make_sandbox_job("alice", "alice running");
+        alice_job.status = "running".to_string();
+        alice_job.success = None;
+        alice_job.completed_at = None;
+        db.save_sandbox_job(&alice_job).await.unwrap();
+
+        let state = build_state(Some(db), None);
+        let auth = two_user_auth();
+        let app = jobs_router(state, auth);
+
+        // Bob tries to cancel alice's job.
+        let req = Request::builder()
+            .method(Method::POST)
+            .uri(format!("/api/jobs/{}/cancel", alice_job.id))
+            .header("Authorization", "Bearer tok-bob")
+            .body(Body::empty())
+            .unwrap();
+        let resp = app.oneshot(req).await.unwrap();
+        assert_eq!(
+            resp.status(),
+            StatusCode::NOT_FOUND,
+            "bob should not be able to cancel alice's job"
+        );
+    }
+}
+
+// ═══════════════════════════════════════════════════════════════════════
+// Routines Isolation Tests
+// ═══════════════════════════════════════════════════════════════════════
+
+#[cfg(feature = "libsql")]
+mod routines_isolation {
+    use super::*;
+    use crate::channels::web::handlers::routines::{
+        routines_delete_handler, routines_detail_handler, routines_list_handler,
+        routines_summary_handler, routines_toggle_handler,
+    };
+    // RoutineStore methods are accessed through the Database supertrait.
+
+    fn routines_router(state: Arc<GatewayState>, auth: MultiAuthState) -> Router {
+        Router::new()
+            .route("/api/routines", get(routines_list_handler))
+            .route("/api/routines/summary", get(routines_summary_handler))
+            .route("/api/routines/{id}", get(routines_detail_handler))
+            .route("/api/routines/{id}/toggle", post(routines_toggle_handler))
+            .route("/api/routines/{id}", delete(routines_delete_handler))
+            .layer(middleware::from_fn_with_state(auth, auth_middleware))
+            .with_state(state)
+    }
+
+    #[tokio::test]
+    async fn test_routines_isolation() {
+        let (db, _dir) = test_db().await;
+
+        // Create routines for alice and bob.
+        let alice_routine = make_routine("alice", "alice-daily");
+        let bob_routine = make_routine("bob", "bob-daily");
+        db.create_routine(&alice_routine).await.unwrap();
+        db.create_routine(&bob_routine).await.unwrap();
+
+        let state = build_state(Some(db), None);
+        let auth = two_user_auth();
+        let app = routines_router(state, auth);
+
+        // Alice sees only her routine in the list.
+        let req = Request::builder()
+            .uri("/api/routines")
+            .header("Authorization", "Bearer tok-alice")
+            .body(Body::empty())
+            .unwrap();
+        let resp = app.clone().oneshot(req).await.unwrap();
+        assert_eq!(resp.status(), StatusCode::OK);
+        let body: serde_json::Value =
+            serde_json::from_slice(&axum::body::to_bytes(resp.into_body(), 8192).await.unwrap())
+                .unwrap();
+        let routines = body["routines"].as_array().unwrap();
+        assert_eq!(routines.len(), 1, "alice should see only her routines");
+        assert_eq!(routines[0]["name"], "alice-daily");
+
+        // Bob sees only his routine.
+        let req = Request::builder()
+            .uri("/api/routines")
+            .header("Authorization", "Bearer tok-bob")
+            .body(Body::empty())
+            .unwrap();
+        let resp = app.clone().oneshot(req).await.unwrap();
+        assert_eq!(resp.status(), StatusCode::OK);
+        let body: serde_json::Value =
+            serde_json::from_slice(&axum::body::to_bytes(resp.into_body(), 8192).await.unwrap())
+                .unwrap();
+        let routines = body["routines"].as_array().unwrap();
+        assert_eq!(routines.len(), 1, "bob should see only his routines");
+        assert_eq!(routines[0]["name"], "bob-daily");
+
+        // Bob cannot view alice's routine detail.
+        let req = Request::builder()
+            .uri(format!("/api/routines/{}", alice_routine.id))
+            .header("Authorization", "Bearer tok-bob")
+            .body(Body::empty())
+            .unwrap();
+        let resp = app.clone().oneshot(req).await.unwrap();
+        assert_eq!(
+            resp.status(),
+            StatusCode::NOT_FOUND,
+            "bob should not see alice's routine detail"
+        );
+
+        // Bob cannot toggle alice's routine.
+        let req = Request::builder()
+            .method(Method::POST)
+            .uri(format!("/api/routines/{}/toggle", alice_routine.id))
+            .header("Authorization", "Bearer tok-bob")
+            .body(Body::empty())
+            .unwrap();
+        let resp = app.clone().oneshot(req).await.unwrap();
+        assert_eq!(
+            resp.status(),
+            StatusCode::NOT_FOUND,
+            "bob should not toggle alice's routine"
+        );
+
+        // Bob cannot delete alice's routine.
+        let req = Request::builder()
+            .method(Method::DELETE)
+            .uri(format!("/api/routines/{}", alice_routine.id))
+            .header("Authorization", "Bearer tok-bob")
+            .body(Body::empty())
+            .unwrap();
+        let resp = app.oneshot(req).await.unwrap();
+        assert_eq!(
+            resp.status(),
+            StatusCode::NOT_FOUND,
+            "bob should not delete alice's routine"
+        );
+    }
+}
+
+// ═══════════════════════════════════════════════════════════════════════
+// Handler Auth Enforcement Tests
+// ═══════════════════════════════════════════════════════════════════════
+
+mod auth_enforcement {
+    use super::*;
+
+    /// Dummy handler that extracts `AuthenticatedUser` — if the auth middleware
+    /// rejects the request, this handler is never reached.
+    async fn authed_handler(AuthenticatedUser(_user): AuthenticatedUser) -> &'static str {
+        "ok"
+    }
+
+    /// Build a router with the real auth middleware and dummy handlers at all
+    /// the paths we want to verify require authentication.
+    fn auth_test_router(auth: MultiAuthState) -> Router {
+        let state = build_state(None, None);
+        Router::new()
+            // Routines
+            .route("/api/routines", get(authed_handler))
+            .route("/api/routines/summary", get(authed_handler))
+            .route("/api/routines/{id}", get(authed_handler))
+            .route("/api/routines/{id}/toggle", post(authed_handler))
+            .route("/api/routines/{id}", delete(authed_handler))
+            // Skills
+            .route("/api/skills", get(authed_handler))
+            .route("/api/skills/search", post(authed_handler))
+            .route("/api/skills/install", post(authed_handler))
+            .route("/api/skills/{name}", delete(authed_handler))
+            // Logs
+            .route("/api/logs/events", get(authed_handler))
+            .route("/api/logs/level", get(authed_handler).put(authed_handler))
+            // Gateway status
+            .route("/api/gateway/status", get(authed_handler))
+            .layer(middleware::from_fn_with_state(auth, auth_middleware))
+            .with_state(state)
+    }
+
+    /// Send a request without auth and assert it returns UNAUTHORIZED.
+    async fn assert_requires_auth(app: &Router, method: Method, uri: &str) {
+        let req = Request::builder()
+            .method(method.clone())
+            .uri(uri)
+            .body(Body::empty())
+            .unwrap();
+        let resp = app.clone().oneshot(req).await.unwrap();
+        assert_eq!(
+            resp.status(),
+            StatusCode::UNAUTHORIZED,
+            "{} {} should require auth",
+            method,
+            uri
+        );
+    }
+
+    /// Send a request with a valid token and assert it succeeds.
+    async fn assert_passes_with_token(app: &Router, method: Method, uri: &str, token: &str) {
+        let req = Request::builder()
+            .method(method.clone())
+            .uri(uri)
+            .header("Authorization", format!("Bearer {token}"))
+            .body(Body::empty())
+            .unwrap();
+        let resp = app.clone().oneshot(req).await.unwrap();
+        assert_eq!(
+            resp.status(),
+            StatusCode::OK,
+            "{} {} should pass with valid token",
+            method,
+            uri
+        );
+    }
+
+    #[tokio::test]
+    async fn test_routines_handlers_require_auth() {
+        let auth = MultiAuthState::single("secret-tok".to_string(), "user".to_string());
+        let app = auth_test_router(auth);
+        let id = Uuid::new_v4();
+
+        assert_requires_auth(&app, Method::GET, "/api/routines").await;
+        assert_requires_auth(&app, Method::GET, "/api/routines/summary").await;
+        assert_requires_auth(&app, Method::GET, &format!("/api/routines/{id}")).await;
+        assert_requires_auth(&app, Method::POST, &format!("/api/routines/{id}/toggle")).await;
+        assert_requires_auth(&app, Method::DELETE, &format!("/api/routines/{id}")).await;
+    }
+
+    #[tokio::test]
+    async fn test_skills_handlers_require_auth() {
+        let auth = MultiAuthState::single("secret-tok".to_string(), "user".to_string());
+        let app = auth_test_router(auth);
+
+        assert_requires_auth(&app, Method::GET, "/api/skills").await;
+        assert_requires_auth(&app, Method::POST, "/api/skills/search").await;
+        assert_requires_auth(&app, Method::POST, "/api/skills/install").await;
+        assert_requires_auth(&app, Method::DELETE, "/api/skills/test-skill").await;
+    }
+
+    #[tokio::test]
+    async fn test_logs_handlers_require_auth() {
+        let auth = MultiAuthState::single("secret-tok".to_string(), "user".to_string());
+        let app = auth_test_router(auth);
+
+        assert_requires_auth(&app, Method::GET, "/api/logs/events").await;
+        assert_requires_auth(&app, Method::GET, "/api/logs/level").await;
+        assert_requires_auth(&app, Method::PUT, "/api/logs/level").await;
+    }
+
+    #[tokio::test]
+    async fn test_gateway_status_requires_auth() {
+        let auth = MultiAuthState::single("secret-tok".to_string(), "user".to_string());
+        let app = auth_test_router(auth);
+
+        assert_requires_auth(&app, Method::GET, "/api/gateway/status").await;
+    }
+
+    #[tokio::test]
+    async fn test_valid_token_passes_all_endpoints() {
+        let auth = MultiAuthState::single("secret-tok".to_string(), "user".to_string());
+        let app = auth_test_router(auth);
+        let id = Uuid::new_v4();
+
+        assert_passes_with_token(&app, Method::GET, "/api/routines", "secret-tok").await;
+        assert_passes_with_token(&app, Method::GET, "/api/skills", "secret-tok").await;
+        assert_passes_with_token(&app, Method::GET, "/api/logs/events", "secret-tok").await;
+        assert_passes_with_token(&app, Method::GET, "/api/gateway/status", "secret-tok").await;
+        assert_passes_with_token(
+            &app,
+            Method::GET,
+            &format!("/api/routines/{id}"),
+            "secret-tok",
+        )
+        .await;
+    }
+
+    #[tokio::test]
+    async fn test_wrong_token_rejected_on_all_endpoints() {
+        let auth = MultiAuthState::single("secret-tok".to_string(), "user".to_string());
+        let app = auth_test_router(auth);
+
+        // Wrong token should be rejected.
+        let req = Request::builder()
+            .uri("/api/routines")
+            .header("Authorization", "Bearer wrong-tok")
+            .body(Body::empty())
+            .unwrap();
+        let resp = app.clone().oneshot(req).await.unwrap();
+        assert_eq!(resp.status(), StatusCode::UNAUTHORIZED);
+
+        let req = Request::builder()
+            .uri("/api/gateway/status")
+            .header("Authorization", "Bearer wrong-tok")
+            .body(Body::empty())
+            .unwrap();
+        let resp = app.oneshot(req).await.unwrap();
+        assert_eq!(resp.status(), StatusCode::UNAUTHORIZED);
+    }
+}

src/channels/web/types.rs

@@ -63,6 +63,9 @@ pub struct TurnInfo {
     pub started_at: String,
     pub completed_at: Option<String>,
     pub tool_calls: Vec<ToolCallInfo>,
+    /// Agent's reasoning narrative for this turn.
+    #[serde(skip_serializing_if = "Option::is_none")]
+    pub narrative: Option<String>,
 }
 
 #[derive(Debug, Serialize)]
@@ -74,6 +77,9 @@ pub struct ToolCallInfo {
     pub result_preview: Option<String>,
     #[serde(skip_serializing_if = "Option::is_none")]
     pub error: Option<String>,
+    /// Agent's reasoning for choosing this tool.
+    #[serde(skip_serializing_if = "Option::is_none")]
+    pub rationale: Option<String>,
 }
 
 #[derive(Debug, Serialize)]
@@ -114,165 +120,9 @@ pub struct ApprovalRequest {
     pub thread_id: Option<String>,
 }
 
-// --- SSE Event Types ---
+// --- App Event (re-exported from ironclaw_common) ---
 
-#[derive(Debug, Clone, Serialize)]
+pub use ironclaw_common::{AppEvent, ToolDecisionDto};
-#[serde(tag = "type")]
-pub enum SseEvent {
-    #[serde(rename = "response")]
-    Response { content: String, thread_id: String },
-    #[serde(rename = "thinking")]
-    Thinking {
-        message: String,
-        #[serde(skip_serializing_if = "Option::is_none")]
-        thread_id: Option<String>,
-    },
-    #[serde(rename = "tool_started")]
-    ToolStarted {
-        name: String,
-        #[serde(skip_serializing_if = "Option::is_none")]
-        thread_id: Option<String>,
-    },
-    #[serde(rename = "tool_completed")]
-    ToolCompleted {
-        name: String,
-        success: bool,
-        #[serde(skip_serializing_if = "Option::is_none")]
-        error: Option<String>,
-        #[serde(skip_serializing_if = "Option::is_none")]
-        parameters: Option<String>,
-        #[serde(skip_serializing_if = "Option::is_none")]
-        thread_id: Option<String>,
-    },
-    #[serde(rename = "tool_result")]
-    ToolResult {
-        name: String,
-        preview: String,
-        #[serde(skip_serializing_if = "Option::is_none")]
-        thread_id: Option<String>,
-    },
-    #[serde(rename = "stream_chunk")]
-    StreamChunk {
-        content: String,
-        #[serde(skip_serializing_if = "Option::is_none")]
-        thread_id: Option<String>,
-    },
-    #[serde(rename = "status")]
-    Status {
-        message: String,
-        #[serde(skip_serializing_if = "Option::is_none")]
-        thread_id: Option<String>,
-    },
-    #[serde(rename = "job_started")]
-    JobStarted {
-        job_id: String,
-        title: String,
-        browse_url: String,
-    },
-    #[serde(rename = "approval_needed")]
-    ApprovalNeeded {
-        request_id: String,
-        tool_name: String,
-        description: String,
-        parameters: String,
-        #[serde(skip_serializing_if = "Option::is_none")]
-        thread_id: Option<String>,
-        /// Whether the "always" auto-approve option should be shown.
-        allow_always: bool,
-    },
-    #[serde(rename = "auth_required")]
-    AuthRequired {
-        extension_name: String,
-        #[serde(skip_serializing_if = "Option::is_none")]
-        instructions: Option<String>,
-        #[serde(skip_serializing_if = "Option::is_none")]
-        auth_url: Option<String>,
-        #[serde(skip_serializing_if = "Option::is_none")]
-        setup_url: Option<String>,
-    },
-    #[serde(rename = "auth_completed")]
-    AuthCompleted {
-        extension_name: String,
-        success: bool,
-        message: String,
-    },
-    #[serde(rename = "error")]
-    Error {
-        message: String,
-        #[serde(skip_serializing_if = "Option::is_none")]
-        thread_id: Option<String>,
-    },
-    #[serde(rename = "heartbeat")]
-    Heartbeat,
-
-    // Sandbox job streaming events (worker + Claude Code bridge)
-    #[serde(rename = "job_message")]
-    JobMessage {
-        job_id: String,
-        role: String,
-        content: String,
-    },
-    #[serde(rename = "job_tool_use")]
-    JobToolUse {
-        job_id: String,
-        tool_name: String,
-        input: serde_json::Value,
-    },
-    #[serde(rename = "job_tool_result")]
-    JobToolResult {
-        job_id: String,
-        tool_name: String,
-        output: String,
-    },
-    #[serde(rename = "job_status")]
-    JobStatus { job_id: String, message: String },
-    #[serde(rename = "job_result")]
-    JobResult {
-        job_id: String,
-        status: String,
-        #[serde(skip_serializing_if = "Option::is_none")]
-        session_id: Option<String>,
-        #[serde(skip_serializing_if = "Option::is_none")]
-        fallback_deliverable: Option<serde_json::Value>,
-    },
-
-    /// An image was generated by a tool.
-    #[serde(rename = "image_generated")]
-    ImageGenerated {
-        data_url: String,
-        #[serde(skip_serializing_if = "Option::is_none")]
-        path: Option<String>,
-        #[serde(skip_serializing_if = "Option::is_none")]
-        thread_id: Option<String>,
-    },
-
-    /// Suggested follow-up messages for the user.
-    #[serde(rename = "suggestions")]
-    Suggestions {
-        suggestions: Vec<String>,
-        #[serde(skip_serializing_if = "Option::is_none")]
-        thread_id: Option<String>,
-    },
-
-    /// Per-turn token usage and cost summary.
-    #[serde(rename = "turn_cost")]
-    TurnCost {
-        input_tokens: u64,
-        output_tokens: u64,
-        cost_usd: String,
-        #[serde(skip_serializing_if = "Option::is_none")]
-        thread_id: Option<String>,
-    },
-
-    /// Extension activation status change (WASM channels).
-    #[serde(rename = "extension_status")]
-    ExtensionStatus {
-        extension_name: String,
-        status: String,
-        #[serde(skip_serializing_if = "Option::is_none")]
-        message: Option<String>,
-    },
-}
 
 // --- Memory ---
 
@@ -784,32 +634,9 @@ pub enum WsServerMessage {
 }
 
 impl WsServerMessage {
-    /// Create a WsServerMessage from an SseEvent.
+    /// Create a WsServerMessage from an AppEvent.
-    pub fn from_sse_event(event: &SseEvent) -> Self {
+    pub fn from_app_event(event: &AppEvent) -> Self {
-        let event_type = match event {
+        let event_type = event.event_type();
-            SseEvent::Response { .. } => "response",
-            SseEvent::Thinking { .. } => "thinking",
-            SseEvent::ToolStarted { .. } => "tool_started",
-            SseEvent::ToolCompleted { .. } => "tool_completed",
-            SseEvent::ToolResult { .. } => "tool_result",
-            SseEvent::StreamChunk { .. } => "stream_chunk",
-            SseEvent::Status { .. } => "status",
-            SseEvent::JobStarted { .. } => "job_started",
-            SseEvent::ApprovalNeeded { .. } => "approval_needed",
-            SseEvent::AuthRequired { .. } => "auth_required",
-            SseEvent::AuthCompleted { .. } => "auth_completed",
-            SseEvent::Error { .. } => "error",
-            SseEvent::Heartbeat => "heartbeat",
-            SseEvent::JobMessage { .. } => "job_message",
-            SseEvent::JobToolUse { .. } => "job_tool_use",
-            SseEvent::JobToolResult { .. } => "job_tool_result",
-            SseEvent::JobStatus { .. } => "job_status",
-            SseEvent::JobResult { .. } => "job_result",
-            SseEvent::ImageGenerated { .. } => "image_generated",
-            SseEvent::Suggestions { .. } => "suggestions",
-            SseEvent::TurnCost { .. } => "turn_cost",
-            SseEvent::ExtensionStatus { .. } => "extension_status",
-        };
         let data = serde_json::to_value(event).unwrap_or(serde_json::Value::Null);
         WsServerMessage::Event {
             event_type: event_type.to_string(),
@@ -1101,12 +928,12 @@ mod tests {
     }
 
     #[test]
-    fn test_ws_server_from_sse_response() {
+    fn test_ws_server_from_app_event_response() {
-        let sse = SseEvent::Response {
+        let event = AppEvent::Response {
             content: "hello".to_string(),
             thread_id: "t1".to_string(),
         };
-        let ws = WsServerMessage::from_sse_event(&sse);
+        let ws = WsServerMessage::from_app_event(&event);
         match ws {
             WsServerMessage::Event { event_type, data } => {
                 assert_eq!(event_type, "response");
@@ -1118,12 +945,12 @@ mod tests {
     }
 
     #[test]
-    fn test_ws_server_from_sse_thinking() {
+    fn test_ws_server_from_app_event_thinking() {
-        let sse = SseEvent::Thinking {
+        let event = AppEvent::Thinking {
             message: "reasoning...".to_string(),
             thread_id: None,
         };
-        let ws = WsServerMessage::from_sse_event(&sse);
+        let ws = WsServerMessage::from_app_event(&event);
         match ws {
             WsServerMessage::Event { event_type, data } => {
                 assert_eq!(event_type, "thinking");
@@ -1134,8 +961,8 @@ mod tests {
     }
 
     #[test]
-    fn test_ws_server_from_sse_approval_needed() {
+    fn test_ws_server_from_app_event_approval_needed() {
-        let sse = SseEvent::ApprovalNeeded {
+        let event = AppEvent::ApprovalNeeded {
             request_id: "r1".to_string(),
             tool_name: "shell".to_string(),
             description: "Run ls".to_string(),
@@ -1143,7 +970,7 @@ mod tests {
             thread_id: Some("t1".to_string()),
             allow_always: true,
         };
-        let ws = WsServerMessage::from_sse_event(&sse);
+        let ws = WsServerMessage::from_app_event(&event);
         match ws {
             WsServerMessage::Event { event_type, data } => {
                 assert_eq!(event_type, "approval_needed");
@@ -1155,9 +982,9 @@ mod tests {
     }
 
     #[test]
-    fn test_ws_server_from_sse_heartbeat() {
+    fn test_ws_server_from_app_event_heartbeat() {
-        let sse = SseEvent::Heartbeat;
+        let event = AppEvent::Heartbeat;
-        let ws = WsServerMessage::from_sse_event(&sse);
+        let ws = WsServerMessage::from_app_event(&event);
         match ws {
             WsServerMessage::Event { event_type, .. } => {
                 assert_eq!(event_type, "heartbeat");
@@ -1197,8 +1024,8 @@ mod tests {
     }
 
     #[test]
-    fn test_sse_auth_required_serialize() {
+    fn test_app_event_auth_required_serialize() {
-        let event = SseEvent::AuthRequired {
+        let event = AppEvent::AuthRequired {
             extension_name: "notion".to_string(),
             instructions: Some("Get your token from...".to_string()),
             auth_url: None,
@@ -1214,8 +1041,8 @@ mod tests {
     }
 
     #[test]
-    fn test_sse_auth_completed_serialize() {
+    fn test_app_event_auth_completed_serialize() {
-        let event = SseEvent::AuthCompleted {
+        let event = AppEvent::AuthCompleted {
             extension_name: "notion".to_string(),
             success: true,
             message: "notion authenticated (3 tools loaded)".to_string(),
@@ -1228,14 +1055,14 @@ mod tests {
     }
 
     #[test]
-    fn test_ws_server_from_sse_auth_required() {
+    fn test_ws_server_from_app_event_auth_required() {
-        let sse = SseEvent::AuthRequired {
+        let event = AppEvent::AuthRequired {
             extension_name: "openai".to_string(),
             instructions: Some("Enter API key".to_string()),
             auth_url: None,
             setup_url: None,
         };
-        let ws = WsServerMessage::from_sse_event(&sse);
+        let ws = WsServerMessage::from_app_event(&event);
         match ws {
             WsServerMessage::Event { event_type, data } => {
                 assert_eq!(event_type, "auth_required");
@@ -1246,13 +1073,13 @@ mod tests {
     }
 
     #[test]
-    fn test_ws_server_from_sse_auth_completed() {
+    fn test_ws_server_from_app_event_auth_completed() {
-        let sse = SseEvent::AuthCompleted {
+        let event = AppEvent::AuthCompleted {
             extension_name: "slack".to_string(),
             success: false,
             message: "Invalid token".to_string(),
         };
-        let ws = WsServerMessage::from_sse_event(&sse);
+        let ws = WsServerMessage::from_app_event(&event);
         match ws {
             WsServerMessage::Event { event_type, data } => {
                 assert_eq!(event_type, "auth_completed");

src/channels/web/util.rs

@@ -2,28 +2,21 @@
 
 use crate::channels::web::types::{ToolCallInfo, TurnInfo};
 
-/// Truncate a string to at most `max_bytes` bytes at a char boundary, appending "...".
+pub use ironclaw_common::truncate_preview;
-///
-/// If the input is wrapped in `<tool_output …>…</tool_output>` and truncation
-/// removes the closing tag, the tag is re-appended so downstream XML parsers
-/// never see an unclosed element.
-pub fn truncate_preview(s: &str, max_bytes: usize) -> String {
-    if s.len() <= max_bytes {
-        return s.to_string();
-    }
-    // Walk backwards from max_bytes to find a valid char boundary
-    let mut end = max_bytes;
-    while end > 0 && !s.is_char_boundary(end) {
-        end -= 1;
-    }
-    let mut result = format!("{}...", &s[..end]);
 
-    // Re-close <tool_output> if truncation cut through the closing tag.
+/// Parse tool call summary JSON objects into `ToolCallInfo` structs.
-    if s.starts_with("<tool_output") && !result.ends_with("</tool_output>") {
+fn parse_tool_call_infos(calls: &[serde_json::Value]) -> Vec<ToolCallInfo> {
-        result.push_str("\n</tool_output>");
+    calls
-    }
+        .iter()
-
+        .map(|c| ToolCallInfo {
-    result
+            name: c["name"].as_str().unwrap_or("unknown").to_string(),
+            has_result: c.get("result_preview").is_some_and(|v| !v.is_null()),
+            has_error: c.get("error").is_some_and(|v| !v.is_null()),
+            result_preview: c["result_preview"].as_str().map(String::from),
+            error: c["error"].as_str().map(String::from),
+            rationale: c["rationale"].as_str().map(String::from),
+        })
+        .collect()
 }
 
 /// Build TurnInfo pairs from flat DB messages (user/tool_calls/assistant triples).
@@ -49,6 +42,7 @@ pub fn build_turns_from_db_messages(
                 started_at: msg.created_at.to_rfc3339(),
                 completed_at: None,
                 tool_calls: Vec::new(),
+                narrative: None,
             };
 
             // Check if next message is a tool_calls record
@@ -56,18 +50,28 @@ pub fn build_turns_from_db_messages(
                 && next.role == "tool_calls"
             {
                 let tc_msg = iter.next().expect("peeked");
-                match serde_json::from_str::<Vec<serde_json::Value>>(&tc_msg.content) {
+                // Parse tool_calls JSON — supports two formats:
-                    Ok(calls) => {
+                // safety: no byte-index slicing; comment describes JSON shape
-                        turn.tool_calls = calls
+                match serde_json::from_str::<serde_json::Value>(&tc_msg.content) {
-                            .iter()
+                    Ok(serde_json::Value::Array(calls)) => {
-                            .map(|c| ToolCallInfo {
+                        // Old format: plain array
-                                name: c["name"].as_str().unwrap_or("unknown").to_string(),
+                        turn.tool_calls = parse_tool_call_infos(&calls);
-                                has_result: c.get("result_preview").is_some(),
+                    }
-                                has_error: c.get("error").is_some(),
+                    Ok(serde_json::Value::Object(obj)) => {
-                                result_preview: c["result_preview"].as_str().map(String::from),
+                        // New wrapped format with narrative
-                                error: c["error"].as_str().map(String::from),
+                        turn.narrative = obj
-                            })
+                            .get("narrative")
-                            .collect();
+                            .and_then(|v| v.as_str())
+                            .map(String::from);
+                        if let Some(serde_json::Value::Array(calls)) = obj.get("calls") {
+                            turn.tool_calls = parse_tool_call_infos(calls);
+                        }
+                    }
+                    Ok(_) => {
+                        tracing::warn!(
+                            message_id = %tc_msg.id,
+                            "Unexpected tool_calls JSON shape in DB, skipping"
+                        );
                     }
                     Err(e) => {
                         tracing::warn!(
@@ -105,6 +109,7 @@ pub fn build_turns_from_db_messages(
                 started_at: msg.created_at.to_rfc3339(),
                 completed_at: Some(msg.created_at.to_rfc3339()),
                 tool_calls: Vec::new(),
+                narrative: None,
             });
             turn_number += 1;
         }
@@ -118,88 +123,6 @@ mod tests {
     use super::*;
     use uuid::Uuid;
 
-    // ---- truncate_preview tests ----
-
-    #[test]
-    fn test_truncate_preview_short_string() {
-        assert_eq!(truncate_preview("hello", 10), "hello");
-    }
-
-    #[test]
-    fn test_truncate_preview_exact_boundary() {
-        assert_eq!(truncate_preview("hello", 5), "hello");
-    }
-
-    #[test]
-    fn test_truncate_preview_truncates_ascii() {
-        assert_eq!(truncate_preview("hello world", 5), "hello...");
-    }
-
-    #[test]
-    fn test_truncate_preview_empty_string() {
-        assert_eq!(truncate_preview("", 10), "");
-    }
-
-    #[test]
-    fn test_truncate_preview_multibyte_char_boundary() {
-        // '€' is 3 bytes (E2 82 AC). "a€b" = [61, E2, 82, AC, 62] = 5 bytes
-        // Truncating at max_bytes=3 should not split the euro sign.
-        let s = "a€b";
-        let result = truncate_preview(s, 3);
-        // max_bytes=3 lands mid-€, so it walks back to byte 1 ("a")
-        assert_eq!(result, "a...");
-    }
-
-    #[test]
-    fn test_truncate_preview_emoji() {
-        // '🦀' is 4 bytes. "hi🦀" = 6 bytes
-        let s = "hi🦀";
-        let result = truncate_preview(s, 4);
-        // max_bytes=4 lands mid-🦀, walks back to byte 2 ("hi")
-        assert_eq!(result, "hi...");
-    }
-
-    #[test]
-    fn test_truncate_preview_cjk() {
-        // CJK characters are 3 bytes each. "你好世界" = 12 bytes
-        let s = "你好世界";
-        let result = truncate_preview(s, 7);
-        // max_bytes=7 lands mid-character (byte 7 is inside 世), walks back to 6 ("你好")
-        assert_eq!(result, "你好...");
-    }
-
-    #[test]
-    fn test_truncate_preview_zero_max_bytes() {
-        assert_eq!(truncate_preview("hello", 0), "...");
-    }
-
-    #[test]
-    fn test_truncate_preview_closes_tool_output_tag() {
-        let s = "<tool_output name=\"search\">\nSome very long content here\n</tool_output>";
-        // Truncate so it cuts before the closing tag
-        let result = truncate_preview(s, 60);
-        assert!(result.ends_with("</tool_output>"));
-        assert!(result.contains("..."));
-    }
-
-    #[test]
-    fn test_truncate_preview_no_extra_close_when_intact() {
-        let s = "<tool_output name=\"echo\">\nshort\n</tool_output>";
-        // The string is short enough not to be truncated
-        let result = truncate_preview(s, 500);
-        assert_eq!(result, s);
-        // Should not have a duplicate closing tag
-        assert_eq!(result.matches("</tool_output>").count(), 1);
-    }
-
-    #[test]
-    fn test_truncate_preview_non_xml_unaffected() {
-        let s = "Just a plain long string that gets truncated";
-        let result = truncate_preview(s, 10);
-        assert_eq!(result, "Just a pla...");
-        assert!(!result.contains("</tool_output>"));
-    }
-
     // ---- build_turns_from_db_messages tests ----
 
     fn make_msg(role: &str, content: &str, offset_ms: i64) -> crate::history::ConversationMessage {
@@ -305,4 +228,52 @@ mod tests {
         assert!(turns[0].tool_calls.is_empty());
         assert_eq!(turns[0].state, "Completed");
     }
+
+    #[test]
+    fn test_build_turns_with_wrapped_tool_calls_format() {
+        let tc_json = serde_json::json!({
+            "narrative": "Searching memory for context before proceeding.",
+            "calls": [
+                {"name": "memory_search", "result_preview": "found 3 items", "rationale": "consult prior context"},
+                {"name": "shell", "error": "permission denied"}
+            ]
+        });
+        let messages = vec![
+            make_msg("user", "Find info", 0),
+            make_msg("tool_calls", &tc_json.to_string(), 500),
+            make_msg("assistant", "Here's what I found", 1000),
+        ];
+        let turns = build_turns_from_db_messages(&messages);
+        assert_eq!(turns.len(), 1);
+        assert_eq!(
+            turns[0].narrative.as_deref(),
+            Some("Searching memory for context before proceeding.")
+        );
+        assert_eq!(turns[0].tool_calls.len(), 2);
+        assert_eq!(turns[0].tool_calls[0].name, "memory_search");
+        assert_eq!(
+            turns[0].tool_calls[0].rationale.as_deref(),
+            Some("consult prior context")
+        );
+        assert!(turns[0].tool_calls[0].has_result);
+        assert_eq!(turns[0].tool_calls[1].name, "shell");
+        assert!(turns[0].tool_calls[1].has_error);
+        assert_eq!(turns[0].response.as_deref(), Some("Here's what I found"));
+    }
+
+    #[test]
+    fn test_build_turns_wrapped_format_without_narrative() {
+        let tc_json = serde_json::json!({
+            "calls": [{"name": "echo", "result_preview": "hello"}]
+        });
+        let messages = vec![
+            make_msg("user", "Say hi", 0),
+            make_msg("tool_calls", &tc_json.to_string(), 500),
+            make_msg("assistant", "Done", 1000),
+        ];
+        let turns = build_turns_from_db_messages(&messages);
+        assert_eq!(turns.len(), 1);
+        assert!(turns[0].narrative.is_none());
+        assert_eq!(turns[0].tool_calls.len(), 1);
+    }
 }

src/channels/web/ws.rs

@@ -62,7 +62,11 @@ impl Default for WsConnectionTracker {
 ///
 /// When either task ends (client disconnect or broadcast closed), both are
 /// cleaned up.
-pub async fn handle_ws_connection(socket: WebSocket, state: Arc<GatewayState>) {
+pub async fn handle_ws_connection(
+    socket: WebSocket,
+    state: Arc<GatewayState>,
+    user: crate::channels::web::auth::UserIdentity,
+) {
     let (mut ws_sink, mut ws_stream) = socket.split();
 
     // Track connection
@@ -71,9 +75,9 @@ pub async fn handle_ws_connection(socket: WebSocket, state: Arc<GatewayState>) {
     }
     let tracker_for_drop = state.ws_tracker.clone();
 
-    // Subscribe to broadcast events (same source as SSE).
+    // Subscribe to broadcast events (same source as SSE), scoped to this user.
     // Reject if we've hit the connection limit.
-    let Some(raw_stream) = state.sse.subscribe_raw() else {
+    let Some(raw_stream) = state.sse.subscribe_raw(Some(user.user_id.clone())) else {
         tracing::warn!("WebSocket rejected: too many connections");
         // Decrement the WS tracker we already incremented above.
         if let Some(ref tracker) = tracker_for_drop {
@@ -93,7 +97,7 @@ pub async fn handle_ws_connection(socket: WebSocket, state: Arc<GatewayState>) {
             let msg = tokio::select! {
                 event = event_stream.next() => {
                     match event {
-                        Some(sse_event) => WsServerMessage::from_sse_event(&sse_event),
+                        Some(app_event) => WsServerMessage::from_app_event(&app_event),
                         None => break, // Broadcast channel closed
                     }
                 }
@@ -117,7 +121,7 @@ pub async fn handle_ws_connection(socket: WebSocket, state: Arc<GatewayState>) {
     });
 
     // Receiver task: read client frames and route to agent
-    let user_id = state.user_id.clone();
+    let user_id = user.user_id;
     while let Some(Ok(frame)) = ws_stream.next().await {
         match frame {
             Message::Text(text) => {
@@ -263,11 +267,15 @@ async fn handle_client_message(
             token,
         } => {
             if let Some(ref ext_mgr) = state.extension_manager {
-                match ext_mgr.configure_token(&extension_name, &token).await {
+                match ext_mgr
+                    .configure_token(&extension_name, &token, user_id)
+                    .await
+                {
                     Ok(result) => {
                         if result.verification.is_some() {
-                            state.sse.broadcast(
+                            state.sse.broadcast_for_user(
-                                crate::channels::web::types::SseEvent::AuthRequired {
+                                user_id,
+                                crate::channels::web::types::AppEvent::AuthRequired {
                                     extension_name: extension_name.clone(),
                                     instructions: Some(result.message),
                                     auth_url: None,
@@ -275,9 +283,10 @@ async fn handle_client_message(
                                 },
                             );
                         } else {
-                            crate::channels::web::server::clear_auth_mode(state).await;
+                            crate::channels::web::server::clear_auth_mode(state, user_id).await;
-                            state.sse.broadcast(
+                            state.sse.broadcast_for_user(
-                                crate::channels::web::types::SseEvent::AuthCompleted {
+                                user_id,
+                                crate::channels::web::types::AppEvent::AuthCompleted {
                                     extension_name,
                                     success: true,
                                     message: result.message,
@@ -288,8 +297,9 @@ async fn handle_client_message(
                     Err(e) => {
                         let msg = format!("Auth failed: {}", e);
                         if matches!(e, crate::extensions::ExtensionError::ValidationFailed(_)) {
-                            state.sse.broadcast(
+                            state.sse.broadcast_for_user(
-                                crate::channels::web::types::SseEvent::AuthRequired {
+                                user_id,
+                                crate::channels::web::types::AppEvent::AuthRequired {
                                     extension_name: extension_name.clone(),
                                     instructions: Some(msg.clone()),
                                     auth_url: None,
@@ -311,7 +321,7 @@ async fn handle_client_message(
             }
         }
         WsClientMessage::AuthCancel { .. } => {
-            crate::channels::web::server::clear_auth_mode(state).await;
+            crate::channels::web::server::clear_auth_mode(state, user_id).await;
         }
         WsClientMessage::Ping => {
             let _ = direct_tx.send(WsServerMessage::Pong).await;
@@ -498,8 +508,9 @@ mod tests {
 
         GatewayState {
             msg_tx: tokio::sync::RwLock::new(msg_tx),
-            sse: SseManager::new(),
+            sse: Arc::new(SseManager::new()),
             workspace: None,
+            workspace_pool: None,
             session_manager: None,
             log_broadcaster: None,
             log_level_handle: None,
@@ -509,13 +520,14 @@ mod tests {
             job_manager: None,
             prompt_queue: None,
             scheduler: None,
-            user_id: "test".to_string(),
+            owner_id: "test".to_string(),
+            default_sender_id: "test".to_string(),
             shutdown_tx: tokio::sync::RwLock::new(None),
             ws_tracker: Some(Arc::new(WsConnectionTracker::new())),
             llm_provider: None,
             skill_registry: None,
             skill_catalog: None,
-            chat_rate_limiter: crate::channels::web::server::RateLimiter::new(30, 60),
+            chat_rate_limiter: crate::channels::web::server::PerUserRateLimiter::new(30, 60),
             oauth_rate_limiter: crate::channels::web::server::RateLimiter::new(10, 60),
             webhook_rate_limiter: crate::channels::web::server::RateLimiter::new(10, 60),
             registry_entries: Vec::new(),

src/cli/mod.rs

@@ -25,6 +25,7 @@ pub mod import;
 mod logs;
 mod mcp;
 pub mod memory;
+mod models;
 pub mod oauth_defaults;
 mod pairing;
 mod registry;
@@ -45,6 +46,7 @@ pub use logs::{LogsCommand, run_logs_command};
 pub use mcp::{McpCommand, run_mcp_command};
 pub use memory::MemoryCommand;
 pub use memory::run_memory_command_with_db;
+pub use models::{ModelsCommand, run_models_command};
 pub use pairing::{PairingCommand, run_pairing_command, run_pairing_command_with_store};
 pub use registry::{RegistryCommand, run_registry_command};
 pub use routines::{RoutinesCommand, run_routines_command};
@@ -217,6 +219,14 @@ pub enum Command {
     )]
     Hooks(HooksCommand),
 
+    /// Manage LLM providers and models
+    #[command(
+        subcommand,
+        about = "Manage LLM providers and models",
+        long_about = "List providers, view current configuration, and set active provider/model.\nExamples:\n  ironclaw models list\n  ironclaw models list openai --verbose\n  ironclaw models status\n  ironclaw models set gpt-4o\n  ironclaw models set-provider anthropic --model claude-sonnet-4-6-20250514"
+    )]
+    Models(ModelsCommand),
+
     /// Probe external dependencies and validate configuration
     #[command(
         about = "Run diagnostics",

src/cli/models.rs

@@ -0,0 +1,864 @@
+//! Models management CLI commands.
+//!
+//! Provides subcommands for listing providers, viewing current model
+//! configuration, and setting the active provider/model. Settings are
+//! persisted to both `config.toml` and `~/.ironclaw/.env` so changes
+//! take effect immediately (no DB connection required).
+
+use clap::Subcommand;
+use std::path::Path;
+
+use crate::llm::registry::ProviderRegistry;
+use crate::settings::Settings;
+
+#[derive(Subcommand, Debug, Clone)]
+pub enum ModelsCommand {
+    /// List providers (or available models for a specific provider)
+    List {
+        /// Show only a specific provider (by ID or alias)
+        provider: Option<String>,
+
+        /// Show detailed information (env vars, base URL, protocol)
+        #[arg(short, long)]
+        verbose: bool,
+
+        /// Output as JSON
+        #[arg(long)]
+        json: bool,
+    },
+
+    /// Show current model configuration
+    Status {
+        /// Output as JSON
+        #[arg(long)]
+        json: bool,
+    },
+
+    /// Set the default model
+    Set {
+        /// Model name (e.g., "gpt-5-mini", "claude-sonnet-4-6-20250514")
+        model: String,
+    },
+
+    /// Set the LLM provider
+    SetProvider {
+        /// Provider ID or alias (e.g., "openai", "anthropic", "ollama")
+        provider: String,
+
+        /// Also set the model (defaults to provider's default model)
+        #[arg(long)]
+        model: Option<String>,
+    },
+}
+
+/// Run the models CLI subcommand.
+pub async fn run_models_command(
+    cmd: ModelsCommand,
+    config_path: Option<&Path>,
+) -> anyhow::Result<()> {
+    match cmd {
+        ModelsCommand::List {
+            provider,
+            verbose,
+            json,
+        } => {
+            if let Some(ref id) = provider {
+                cmd_show_provider(id, verbose, json, config_path).await
+            } else {
+                cmd_list_providers(verbose, json, config_path).await
+            }
+        }
+        ModelsCommand::Status { json } => cmd_status(json, config_path),
+        ModelsCommand::Set { model } => cmd_set_model(&model, config_path),
+        ModelsCommand::SetProvider { provider, model } => {
+            cmd_set_provider(&provider, model.as_deref(), config_path)
+        }
+    }
+}
+
+// ─── Shared helpers ───────────────────────────────────────────────
+
+/// Resolve the currently active backend and model from env + settings.
+fn resolve_active(config_path: Option<&Path>) -> (String, String) {
+    let settings = load_settings(config_path);
+    resolve_active_from_settings(&settings)
+}
+
+/// Resolve active backend + model from a pre-loaded Settings.
+fn resolve_active_from_settings(settings: &Settings) -> (String, String) {
+    let backend = std::env::var("LLM_BACKEND")
+        .ok()
+        .or_else(|| settings.llm_backend.clone())
+        .unwrap_or_else(|| "nearai".to_string());
+
+    let registry = ProviderRegistry::load();
+
+    let canonical_backend = registry
+        .find(&backend)
+        .map(|d| d.id.clone())
+        .unwrap_or_else(|| backend.clone());
+
+    let model = if canonical_backend == "nearai" {
+        std::env::var("NEARAI_MODEL")
+            .ok()
+            .or_else(|| settings.selected_model.clone())
+            .unwrap_or_else(|| "qwen2.5-72b-instruct:free".to_string())
+    } else if let Some(def) = registry.find(&canonical_backend) {
+        std::env::var(&def.model_env)
+            .ok()
+            .or_else(|| settings.selected_model.clone())
+            .unwrap_or_else(|| def.default_model.clone())
+    } else {
+        settings
+            .selected_model
+            .clone()
+            .unwrap_or_else(|| "unknown".to_string())
+    };
+
+    (canonical_backend, model)
+}
+
+fn load_settings(config_path: Option<&Path>) -> Settings {
+    if let Some(path) = config_path {
+        Settings::load_toml(path).ok().flatten().unwrap_or_default()
+    } else {
+        let toml_path = config_toml_path();
+        if toml_path.exists() {
+            Settings::load_toml(&toml_path)
+                .ok()
+                .flatten()
+                .unwrap_or_default()
+        } else {
+            Settings::load()
+        }
+    }
+}
+
+fn save_settings(settings: &Settings, config_path: Option<&Path>) -> anyhow::Result<()> {
+    let path = config_path
+        .map(|p| p.to_path_buf())
+        .unwrap_or_else(config_toml_path);
+
+    settings
+        .save_toml(&path)
+        .map_err(|e| anyhow::anyhow!("{}", e))?;
+
+    Ok(())
+}
+
+fn config_toml_path() -> std::path::PathBuf {
+    crate::bootstrap::ironclaw_base_dir().join("config.toml")
+}
+
+/// Try to fetch the live model list from a provider.
+///
+/// Best-effort: returns `None` if config loading, provider creation, or the
+/// `list_models()` call fails (missing API key, network error, etc.).
+async fn try_fetch_models(provider_id: &str, config_path: Option<&Path>) -> Option<Vec<String>> {
+    let config = crate::config::Config::from_env_with_toml(config_path)
+        .await
+        .ok()?;
+
+    // Override backend to the requested provider so create_llm_provider
+    // constructs the right one.
+    let mut llm_config = config.llm.clone();
+    llm_config.backend = provider_id.to_string();
+
+    // For registry providers, resolve the RegistryProviderConfig if not
+    // already set for this backend.
+    if provider_id != "nearai" && provider_id != "bedrock" {
+        let registry = ProviderRegistry::load();
+        if let Some(def) = registry.find(provider_id)
+            && llm_config
+                .provider
+                .as_ref()
+                .is_none_or(|p| p.provider_id != def.id)
+        {
+            // Build a minimal RegistryProviderConfig from env + registry
+            let api_key = def
+                .api_key_env
+                .as_ref()
+                .and_then(|env| std::env::var(env).ok());
+            if def.api_key_required && api_key.is_none() {
+                return None;
+            }
+            let base_url = def.default_base_url.clone().unwrap_or_default();
+            llm_config.provider = Some(crate::llm::RegistryProviderConfig {
+                protocol: def.protocol,
+                provider_id: def.id.clone(),
+                model: def.default_model.clone(),
+                api_key: api_key.map(secrecy::SecretString::from),
+                base_url,
+                extra_headers: Vec::new(),
+                oauth_token: None,
+                is_codex_chatgpt: false,
+                refresh_token: None,
+                auth_path: None,
+                cache_retention: Default::default(),
+                unsupported_params: def.unsupported_params.clone(),
+            });
+        }
+    }
+
+    let session = crate::llm::create_session_manager(config.llm.session.clone()).await;
+    let provider = crate::llm::create_llm_provider(&llm_config, session)
+        .await
+        .ok()?;
+    provider.list_models().await.ok().filter(|m| !m.is_empty())
+}
+
+/// Print available models section (text output).
+fn print_model_list(models: &Option<Vec<String>>, active_model: Option<&String>) {
+    match models {
+        Some(models) => {
+            println!("\n  Available models ({}):", models.len());
+            for m in models {
+                let marker = active_model
+                    .filter(|a| a.as_str() == m)
+                    .map(|_| " (active)")
+                    .unwrap_or("");
+                println!("    {}{}", m, marker);
+            }
+        }
+        None => {
+            println!(
+                "\n  Could not fetch model list (missing credentials or provider unavailable)."
+            );
+        }
+    }
+}
+
+/// Also update `~/.ironclaw/.env` so changes take effect immediately.
+///
+/// Skipped when `config_path` is `Some` (custom `--config`), because the user
+/// is explicitly targeting a different config file and we must not pollute the
+/// default profile's `.env`.
+fn sync_to_dotenv(config_path: Option<&Path>, vars: &[(&str, &str)]) {
+    if config_path.is_some() {
+        return;
+    }
+    if let Err(e) = crate::bootstrap::upsert_bootstrap_vars(vars) {
+        eprintln!("Warning: failed to update .env: {}", e);
+    }
+}
+
+// ─── status ───────────────────────────────────────────────────────
+
+fn cmd_status(json: bool, config_path: Option<&Path>) -> anyhow::Result<()> {
+    let settings = load_settings(config_path);
+    let (backend, model) = resolve_active_from_settings(&settings);
+    let registry = ProviderRegistry::load();
+
+    let fallback = std::env::var("NEARAI_FALLBACK_MODEL").ok();
+    let cheap = std::env::var("NEARAI_CHEAP_MODEL").ok();
+
+    let description = if backend == "nearai" {
+        "NEAR AI inference (default)".to_string()
+    } else {
+        registry
+            .find(&backend)
+            .map(|d| d.description.clone())
+            .unwrap_or_default()
+    };
+
+    if json {
+        let v = serde_json::json!({
+            "provider": backend,
+            "model": model,
+            "description": description,
+            "fallback_model": fallback,
+            "cheap_model": cheap,
+        });
+        println!(
+            "{}",
+            serde_json::to_string_pretty(&v).unwrap_or_else(|_| "{}".to_string())
+        );
+        return Ok(());
+    }
+
+    println!("Provider: {} ({})", backend, description);
+    println!("Model:    {}", model);
+    if let Some(ref fb) = fallback {
+        println!("Fallback: {}", fb);
+    }
+    if let Some(ref ch) = cheap {
+        println!("Cheap:    {}", ch);
+    }
+
+    Ok(())
+}
+
+// ─── set ──────────────────────────────────────────────────────────
+
+fn cmd_set_model(model: &str, config_path: Option<&Path>) -> anyhow::Result<()> {
+    let trimmed = model.trim();
+    if trimmed.is_empty() {
+        anyhow::bail!("Model name cannot be empty");
+    }
+
+    let mut settings = load_settings(config_path);
+    let registry = ProviderRegistry::load();
+
+    // Warn if model name doesn't match any known provider's default model
+    let known_model = registry.all().iter().any(|d| d.default_model == trimmed)
+        || trimmed.contains("qwen")  // nearai models
+        || trimmed.contains("llama")
+        || trimmed.contains("gpt")
+        || trimmed.contains("claude")
+        || trimmed.contains("gemini")
+        || trimmed.contains("mistral");
+    if !known_model {
+        eprintln!(
+            "Warning: '{}' is not a recognized model name. Proceeding anyway.",
+            trimmed
+        );
+    }
+
+    settings.selected_model = Some(trimmed.to_string());
+    save_settings(&settings, config_path)?;
+
+    let backend = std::env::var("LLM_BACKEND")
+        .ok()
+        .or_else(|| settings.llm_backend.clone())
+        .unwrap_or_else(|| "nearai".to_string());
+
+    // Also write to .env so the change takes effect immediately
+    let model_env = if backend == "nearai" {
+        "NEARAI_MODEL".to_string()
+    } else {
+        registry
+            .find(&backend)
+            .map(|d| d.model_env.clone())
+            .unwrap_or_default()
+    };
+    if !model_env.is_empty() {
+        sync_to_dotenv(config_path, &[(&model_env, trimmed)]);
+    }
+
+    println!("Model set to '{}' (provider: {})", trimmed, backend);
+    println!(
+        "Saved to {}",
+        config_path
+            .map(|p| p.display().to_string())
+            .unwrap_or_else(|| config_toml_path().display().to_string())
+    );
+
+    Ok(())
+}
+
+// ─── set-provider ─────────────────────────────────────────────────
+
+fn cmd_set_provider(
+    provider: &str,
+    model: Option<&str>,
+    config_path: Option<&Path>,
+) -> anyhow::Result<()> {
+    let registry = ProviderRegistry::load();
+
+    // Validate and normalize provider
+    let canonical_id = if provider == "nearai" || provider == "near_ai" || provider == "near" {
+        "nearai".to_string()
+    } else {
+        let def = registry.find(provider).ok_or_else(|| {
+            let known: Vec<&str> = std::iter::once("nearai")
+                .chain(registry.all().iter().map(|d| d.id.as_str()))
+                .collect();
+            anyhow::anyhow!(
+                "Unknown provider '{}'. Known providers: {}",
+                provider,
+                known.join(", ")
+            )
+        })?;
+        def.id.clone()
+    };
+
+    // Resolve model: explicit > provider default
+    let resolved_model = if let Some(m) = model {
+        m.to_string()
+    } else if canonical_id == "nearai" {
+        "qwen2.5-72b-instruct:free".to_string()
+    } else if let Some(def) = registry.find(&canonical_id) {
+        def.default_model.clone()
+    } else {
+        "default".to_string()
+    };
+
+    let mut settings = load_settings(config_path);
+    settings.llm_backend = Some(canonical_id.clone());
+    settings.selected_model = Some(resolved_model.clone());
+    save_settings(&settings, config_path)?;
+
+    // Also write to .env so the change takes effect immediately
+    let model_env = if canonical_id == "nearai" {
+        "NEARAI_MODEL".to_string()
+    } else {
+        registry
+            .find(&canonical_id)
+            .map(|d| d.model_env.clone())
+            .unwrap_or_default()
+    };
+    let mut vars: Vec<(&str, &str)> = vec![("LLM_BACKEND", &canonical_id)];
+    if !model_env.is_empty() {
+        vars.push((&model_env, &resolved_model));
+    }
+    sync_to_dotenv(config_path, &vars);
+
+    println!(
+        "Provider set to '{}', model set to '{}'",
+        canonical_id, resolved_model
+    );
+    println!(
+        "Saved to {}",
+        config_path
+            .map(|p| p.display().to_string())
+            .unwrap_or_else(|| config_toml_path().display().to_string())
+    );
+
+    Ok(())
+}
+
+// ─── list ─────────────────────────────────────────────────────────
+
+/// List all providers with their default models.
+async fn cmd_list_providers(
+    verbose: bool,
+    json: bool,
+    config_path: Option<&Path>,
+) -> anyhow::Result<()> {
+    let registry = ProviderRegistry::load();
+    let (active_backend, active_model) = resolve_active(config_path);
+
+    if json {
+        let mut entries: Vec<serde_json::Value> = Vec::new();
+
+        // NEAR AI (not in registry)
+        let nearai_active = active_backend == "nearai";
+        entries.push(serde_json::json!({
+            "id": "nearai",
+            "description": "NEAR AI inference (default)",
+            "default_model": "qwen2.5-72b-instruct:free",
+            "active": nearai_active,
+            "active_model": if nearai_active { Some(&active_model) } else { None },
+        }));
+
+        for def in registry.all() {
+            let is_active = active_backend == def.id;
+            let mut v = serde_json::json!({
+                "id": def.id,
+                "description": def.description,
+                "default_model": def.default_model,
+                "protocol": format!("{:?}", def.protocol),
+                "active": is_active,
+            });
+            if is_active {
+                v["active_model"] = serde_json::json!(active_model);
+            }
+            if verbose {
+                v["aliases"] = serde_json::json!(def.aliases);
+                v["model_env"] = serde_json::json!(def.model_env);
+                v["api_key_env"] = serde_json::json!(def.api_key_env);
+                v["api_key_required"] = serde_json::json!(def.api_key_required);
+                if let Some(ref url) = def.default_base_url {
+                    v["base_url"] = serde_json::json!(url);
+                }
+                if let Some(ref setup) = def.setup {
+                    v["can_list_models"] = serde_json::json!(setup.can_list_models());
+                }
+            }
+            entries.push(v);
+        }
+
+        println!(
+            "{}",
+            serde_json::to_string_pretty(&entries).unwrap_or_else(|_| "[]".to_string())
+        );
+        return Ok(());
+    }
+
+    let providers = registry.all();
+
+    println!("Active: {} (model: {})\n", active_backend, active_model);
+    println!(
+        "{} provider(s) available:\n",
+        providers.len() + 1 // +1 for NEAR AI
+    );
+
+    // NEAR AI (not in registry)
+    let nearai_marker = if active_backend == "nearai" { " *" } else { "" };
+    if verbose {
+        println!("  nearai{}", nearai_marker);
+        println!("    Description:   NEAR AI inference (default)");
+        println!("    Default model: qwen2.5-72b-instruct:free");
+        println!("    Model env:     NEARAI_MODEL");
+        if active_backend == "nearai" {
+            println!("    Active model:  {}", active_model);
+        }
+        println!();
+    } else {
+        println!(
+            "  {:<22} {:<40} NEAR AI inference (default)",
+            format!("nearai{nearai_marker}"),
+            "qwen2.5-72b-instruct:free"
+        );
+    }
+
+    for def in providers {
+        let is_active = active_backend == def.id;
+        let marker = if is_active { " *" } else { "" };
+
+        if verbose {
+            println!("  {}{}", def.id, marker);
+            println!("    Description:   {}", def.description);
+            println!("    Default model: {}", def.default_model);
+            println!("    Protocol:      {:?}", def.protocol);
+            println!("    Model env:     {}", def.model_env);
+            if let Some(ref env) = def.api_key_env {
+                println!(
+                    "    API key env:   {} ({})",
+                    env,
+                    if def.api_key_required {
+                        "required"
+                    } else {
+                        "optional"
+                    }
+                );
+            }
+            if let Some(ref url) = def.default_base_url {
+                println!("    Base URL:      {}", url);
+            }
+            if !def.aliases.is_empty() {
+                println!("    Aliases:       {}", def.aliases.join(", "));
+            }
+            if is_active {
+                println!("    Active model:  {}", active_model);
+            }
+            println!();
+        } else {
+            let model_display = if is_active {
+                active_model.clone()
+            } else {
+                def.default_model.clone()
+            };
+            println!(
+                "  {:<22} {:<40} {}",
+                format!("{}{marker}", def.id),
+                model_display,
+                def.description,
+            );
+        }
+    }
+
+    if !verbose {
+        println!();
+        println!("* = active provider. Use --verbose for details.");
+    }
+
+    Ok(())
+}
+
+/// Show details for a specific provider.
+async fn cmd_show_provider(
+    id: &str,
+    verbose: bool,
+    json: bool,
+    config_path: Option<&Path>,
+) -> anyhow::Result<()> {
+    let registry = ProviderRegistry::load();
+    let (active_backend, active_model) = resolve_active(config_path);
+
+    // Resolve canonical ID for model fetching
+    let canonical_id = if id == "nearai" || id == "near_ai" || id == "near" {
+        "nearai".to_string()
+    } else {
+        registry
+            .find(id)
+            .map(|d| d.id.clone())
+            .unwrap_or_else(|| id.to_string())
+    };
+
+    // Try to fetch live model list from the provider
+    let live_models = try_fetch_models(&canonical_id, config_path).await;
+
+    // Check NEAR AI first (not in registry)
+    if id == "nearai" || id == "near_ai" || id == "near" {
+        let is_active = active_backend == "nearai";
+        if json {
+            let mut v = serde_json::json!({
+                "id": "nearai",
+                "description": "NEAR AI inference (default)",
+                "default_model": "qwen2.5-72b-instruct:free",
+                "model_env": "NEARAI_MODEL",
+                "active": is_active,
+            });
+            if is_active {
+                v["active_model"] = serde_json::json!(active_model);
+            }
+            if let Some(ref models) = live_models {
+                v["available_models"] = serde_json::json!(models);
+            }
+            println!(
+                "{}",
+                serde_json::to_string_pretty(&v).unwrap_or_else(|_| "{}".to_string())
+            );
+        } else {
+            println!("Provider: nearai");
+            println!("  Description:   NEAR AI inference (default)");
+            println!("  Default model: qwen2.5-72b-instruct:free");
+            println!("  Model env:     NEARAI_MODEL");
+            println!("  Active:        {}", if is_active { "yes" } else { "no" });
+            if is_active {
+                println!("  Active model:  {}", active_model);
+            }
+            print_model_list(&live_models, is_active.then_some(&active_model));
+        }
+        return Ok(());
+    }
+
+    let def = registry.find(id).ok_or_else(|| {
+        let known: Vec<&str> = std::iter::once("nearai")
+            .chain(registry.all().iter().map(|d| d.id.as_str()))
+            .collect();
+        anyhow::anyhow!(
+            "Unknown provider '{}'. Known providers: {}",
+            id,
+            known.join(", ")
+        )
+    })?;
+
+    let is_active = active_backend == def.id;
+
+    if json {
+        let mut v = serde_json::json!({
+            "id": def.id,
+            "description": def.description,
+            "protocol": format!("{:?}", def.protocol),
+            "default_model": def.default_model,
+            "model_env": def.model_env,
+            "api_key_env": def.api_key_env,
+            "api_key_required": def.api_key_required,
+            "aliases": def.aliases,
+            "active": is_active,
+        });
+        if let Some(ref url) = def.default_base_url {
+            v["base_url"] = serde_json::json!(url);
+        }
+        if let Some(ref setup) = def.setup {
+            v["can_list_models"] = serde_json::json!(setup.can_list_models());
+            v["display_name"] = serde_json::json!(setup.display_name());
+        }
+        if is_active {
+            v["active_model"] = serde_json::json!(active_model);
+        }
+        if verbose && !def.unsupported_params.is_empty() {
+            v["unsupported_params"] = serde_json::json!(def.unsupported_params);
+        }
+        if let Some(ref models) = live_models {
+            v["available_models"] = serde_json::json!(models);
+        }
+        println!(
+            "{}",
+            serde_json::to_string_pretty(&v).unwrap_or_else(|_| "{}".to_string())
+        );
+        return Ok(());
+    }
+
+    println!("Provider: {}", def.id);
+    println!("  Description:   {}", def.description);
+    println!("  Protocol:      {:?}", def.protocol);
+    println!("  Default model: {}", def.default_model);
+    println!("  Model env:     {}", def.model_env);
+    if let Some(ref env) = def.api_key_env {
+        println!(
+            "  API key env:   {} ({})",
+            env,
+            if def.api_key_required {
+                "required"
+            } else {
+                "optional"
+            }
+        );
+    }
+    if let Some(ref url) = def.default_base_url {
+        println!("  Base URL:      {}", url);
+    }
+    if !def.aliases.is_empty() {
+        println!("  Aliases:       {}", def.aliases.join(", "));
+    }
+    if let Some(ref setup) = def.setup {
+        println!(
+            "  List models:   {}",
+            if setup.can_list_models() {
+                "supported"
+            } else {
+                "not supported"
+            }
+        );
+        println!("  Display name:  {}", setup.display_name());
+    }
+    if !def.unsupported_params.is_empty() {
+        println!("  Unsupported:   {}", def.unsupported_params.join(", "));
+    }
+    println!("  Active:        {}", if is_active { "yes" } else { "no" });
+    if is_active {
+        println!("  Active model:  {}", active_model);
+    }
+    print_model_list(&live_models, is_active.then_some(&active_model));
+
+    Ok(())
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn resolve_active_defaults_to_nearai() {
+        let settings = Settings::default();
+        assert!(settings.llm_backend.is_none());
+        assert!(settings.selected_model.is_none());
+    }
+
+    #[test]
+    fn registry_loads_all_providers() {
+        let registry = ProviderRegistry::load();
+        let all = registry.all();
+        assert!(
+            all.len() >= 10,
+            "should have at least 10 built-in providers, got {}",
+            all.len()
+        );
+    }
+
+    #[test]
+    fn registry_find_by_alias() {
+        let registry = ProviderRegistry::load();
+        let def = registry
+            .find("claude")
+            .expect("claude alias should resolve");
+        assert_eq!(def.id, "anthropic");
+    }
+
+    #[test]
+    fn all_providers_have_description() {
+        let registry = ProviderRegistry::load();
+        for def in registry.all() {
+            assert!(
+                !def.description.is_empty(),
+                "provider {} should have a description",
+                def.id
+            );
+        }
+    }
+
+    #[test]
+    fn set_model_persists_to_toml() {
+        let dir = tempfile::tempdir().expect("create temp dir");
+        let toml_path = dir.path().join("config.toml");
+
+        cmd_set_model("gpt-5-mini", Some(&toml_path)).expect("set model");
+
+        let settings = Settings::load_toml(&toml_path)
+            .expect("read toml")
+            .expect("should have settings");
+        assert_eq!(settings.selected_model.as_deref(), Some("gpt-5-mini"));
+    }
+
+    #[test]
+    fn set_provider_validates_unknown() {
+        let dir = tempfile::tempdir().expect("create temp dir");
+        let toml_path = dir.path().join("config.toml");
+
+        let result = cmd_set_provider("nonexistent_provider", None, Some(&toml_path));
+        assert!(result.is_err());
+        let err = result.unwrap_err().to_string();
+        assert!(
+            err.contains("Unknown provider"),
+            "should mention unknown provider: {}",
+            err
+        );
+    }
+
+    #[test]
+    fn set_provider_persists_to_toml() {
+        let dir = tempfile::tempdir().expect("create temp dir");
+        let toml_path = dir.path().join("config.toml");
+
+        cmd_set_provider("groq", None, Some(&toml_path)).expect("set provider");
+
+        let settings = Settings::load_toml(&toml_path)
+            .expect("read toml")
+            .expect("should have settings");
+        assert_eq!(settings.llm_backend.as_deref(), Some("groq"));
+        assert_eq!(
+            settings.selected_model.as_deref(),
+            Some("llama-3.3-70b-versatile")
+        );
+    }
+
+    #[test]
+    fn set_provider_with_custom_model() {
+        let dir = tempfile::tempdir().expect("create temp dir");
+        let toml_path = dir.path().join("config.toml");
+
+        cmd_set_provider("anthropic", Some("claude-opus-4-6"), Some(&toml_path))
+            .expect("set provider with model");
+
+        let settings = Settings::load_toml(&toml_path)
+            .expect("read toml")
+            .expect("should have settings");
+        assert_eq!(settings.llm_backend.as_deref(), Some("anthropic"));
+        assert_eq!(settings.selected_model.as_deref(), Some("claude-opus-4-6"));
+    }
+
+    #[test]
+    fn custom_config_does_not_pollute_default_dotenv() {
+        let dir = tempfile::tempdir().expect("create temp dir");
+        let toml_path = dir.path().join("config.toml");
+
+        // With a custom config path, sync_to_dotenv should be a no-op
+        // (it returns early when config_path is Some).
+        // We verify by checking that cmd_set_provider succeeds without
+        // trying to write to the default ~/.ironclaw/.env.
+        cmd_set_provider("groq", None, Some(&toml_path)).expect("set provider with custom config");
+
+        let settings = Settings::load_toml(&toml_path)
+            .expect("read toml")
+            .expect("should have settings");
+        assert_eq!(settings.llm_backend.as_deref(), Some("groq"));
+        // The key assertion is that no error was thrown trying to write
+        // to the default .env — sync_to_dotenv skipped it.
+    }
+
+    #[test]
+    fn set_model_rejects_empty_name() {
+        let dir = tempfile::tempdir().expect("create temp dir");
+        let toml_path = dir.path().join("config.toml");
+
+        let result = cmd_set_model("", Some(&toml_path));
+        assert!(result.is_err());
+        assert!(
+            result.unwrap_err().to_string().contains("cannot be empty"),
+            "should reject empty model name"
+        );
+
+        let result2 = cmd_set_model("   ", Some(&toml_path));
+        assert!(result2.is_err());
+    }
+
+    #[test]
+    fn set_provider_normalizes_alias() {
+        let dir = tempfile::tempdir().expect("create temp dir");
+        let toml_path = dir.path().join("config.toml");
+
+        cmd_set_provider("claude", None, Some(&toml_path)).expect("set via alias");
+
+        let settings = Settings::load_toml(&toml_path)
+            .expect("read toml")
+            .expect("should have settings");
+        assert_eq!(
+            settings.llm_backend.as_deref(),
+            Some("anthropic"),
+            "alias should be normalized to canonical ID"
+        );
+    }
+}

src/cli/oauth_defaults.rs

@@ -62,6 +62,30 @@ pub fn builtin_client_id_override_env(secret_name: &str) -> Option<&'static str>
     }
 }
 
+/// Suppress the baked-in desktop OAuth client secret when a hosted proxy is configured.
+///
+/// In hosted deployments, IronClaw may resolve the platform Google client ID from
+/// environment variables while still falling back to the baked-in desktop secret.
+/// That client_id/client_secret mismatch breaks Google token exchange and refresh.
+///
+/// When the proxy is configured, the platform will inject the correct server-side
+/// secret for matching platform credentials, so the baked-in secret must be omitted.
+pub fn hosted_proxy_client_secret(
+    client_secret: &Option<String>,
+    builtin: Option<&OAuthCredentials>,
+    exchange_proxy_configured: bool,
+) -> Option<String> {
+    if !exchange_proxy_configured {
+        return client_secret.clone();
+    }
+
+    let builtin_secret = builtin.map(|credentials| credentials.client_secret);
+    match (client_secret, builtin_secret) {
+        (Some(resolved), Some(baked_in)) if resolved == baked_in => None,
+        _ => client_secret.clone(),
+    }
+}
+
 // ── Shared callback server ──────────────────────────────────────────────
 
 // Core OAuth callback infrastructure is defined in `crate::llm::oauth_helpers`
@@ -447,8 +471,8 @@ pub struct PendingOAuthFlow {
     pub user_id: String,
     /// Secrets store reference for token persistence.
     pub secrets: Arc<dyn SecretsStore + Send + Sync>,
-    /// SSE broadcast sender for notifying the web UI.
+    /// SSE broadcast manager for notifying the web UI.
-    pub sse_sender: Option<tokio::sync::broadcast::Sender<crate::channels::web::types::SseEvent>>,
+    pub sse_manager: Option<Arc<crate::channels::web::sse::SseManager>>,
     /// Gateway auth token for authenticating with the platform token exchange proxy.
     pub gateway_token: Option<String>,
     /// Additional form params for the token exchange request.
@@ -661,6 +685,48 @@ pub struct ProxyTokenExchangeRequest<'a> {
     pub extra_token_params: &'a HashMap<String, String>,
 }
 
+pub struct ProxyRefreshTokenRequest<'a> {
+    pub proxy_url: &'a str,
+    pub gateway_token: &'a str,
+    pub token_url: &'a str,
+    pub client_id: &'a str,
+    pub client_secret: Option<&'a str>,
+    pub refresh_token: &'a str,
+    pub provider: Option<&'a str>,
+}
+
+fn oauth_token_response_from_json(
+    token_data: serde_json::Value,
+    access_token_field: &str,
+) -> Result<OAuthTokenResponse, OAuthCallbackError> {
+    let access_token = token_data
+        .get(access_token_field)
+        .and_then(|v| v.as_str())
+        .ok_or_else(|| {
+            let fields: Vec<&str> = token_data
+                .as_object()
+                .map(|o| o.keys().map(|k| k.as_str()).collect())
+                .unwrap_or_default();
+            OAuthCallbackError::Io(format!(
+                "No '{}' field in proxy response (fields present: {:?})",
+                access_token_field, fields
+            ))
+        })?
+        .to_string();
+
+    let refresh_token = token_data
+        .get("refresh_token")
+        .and_then(|v| v.as_str())
+        .map(String::from);
+    let expires_in = token_data.get("expires_in").and_then(|v| v.as_u64());
+
+    Ok(OAuthTokenResponse {
+        access_token,
+        refresh_token,
+        expires_in,
+    })
+}
+
 /// Exchange an OAuth authorization code via the platform's token exchange proxy.
 ///
 /// Authenticated via the gateway auth token (Bearer header). The caller may
@@ -682,6 +748,7 @@ pub async fn exchange_via_proxy(
 
     let client = reqwest::Client::builder()
         .timeout(Duration::from_secs(60))
+        .redirect(reqwest::redirect::Policy::none())
         .build()
         .map_err(|e| OAuthCallbackError::Io(format!("Failed to build HTTP client: {}", e)))?;
     let mut params = vec![
@@ -724,41 +791,350 @@ pub async fn exchange_via_proxy(
         .json()
         .await
         .map_err(|e| OAuthCallbackError::Io(format!("Failed to parse proxy response: {}", e)))?;
+    oauth_token_response_from_json(token_data, request.access_token_field)
+}
 
-    let access_token = token_data
+/// Refresh an OAuth access token via the platform's token refresh proxy.
-        .get(request.access_token_field)
+///
-        .and_then(|v| v.as_str())
+/// Authenticated via the gateway auth token (Bearer header). The caller may
-        .ok_or_else(|| {
+/// either rely on proxy-side secret lookup or forward a `client_secret` when
-            let fields: Vec<&str> = token_data
+/// the provider requires it.
-                .as_object()
+pub async fn refresh_token_via_proxy(
-                .map(|o| o.keys().map(|k| k.as_str()).collect())
+    request: ProxyRefreshTokenRequest<'_>,
-                .unwrap_or_default();
+) -> Result<OAuthTokenResponse, OAuthCallbackError> {
-            OAuthCallbackError::Io(format!(
+    if request.gateway_token.is_empty() {
-                "No '{}' field in proxy response (fields present: {:?})",
+        return Err(OAuthCallbackError::Io(
-                request.access_token_field, fields
+            "Gateway auth token is required for proxy token refresh".to_string(),
-            ))
+        ));
-        })?
+    }
-        .to_string();
 
-    let refresh_token = token_data
+    let refresh_url = format!("{}/oauth/refresh", request.proxy_url.trim_end_matches('/'));
-        .get("refresh_token")
+    let client = reqwest::Client::builder()
-        .and_then(|v| v.as_str())
+        .timeout(Duration::from_secs(15))
-        .map(String::from);
+        .redirect(reqwest::redirect::Policy::none())
-    let expires_in = token_data.get("expires_in").and_then(|v| v.as_u64());
+        .build()
+        .map_err(|e| OAuthCallbackError::Io(format!("Failed to build HTTP client: {}", e)))?;
 
-    Ok(OAuthTokenResponse {
+    let mut params = vec![
-        access_token,
+        ("refresh_token", request.refresh_token.to_string()),
-        refresh_token,
+        ("token_url", request.token_url.to_string()),
-        expires_in,
+        ("client_id", request.client_id.to_string()),
-    })
+    ];
+    if let Some(secret) = request.client_secret {
+        params.push(("client_secret", secret.to_string()));
+    }
+    if let Some(provider) = request.provider {
+        params.push(("provider", provider.to_string()));
+    }
+
+    let response = client
+        .post(&refresh_url)
+        .bearer_auth(request.gateway_token)
+        .form(&params)
+        .send()
+        .await
+        .map_err(|e| {
+            OAuthCallbackError::Io(format!("Token refresh proxy request failed: {}", e))
+        })?;
+
+    if !response.status().is_success() {
+        let status = response.status();
+        let body = response.text().await.unwrap_or_default();
+        return Err(OAuthCallbackError::Io(format!(
+            "Token refresh proxy failed: {} - {}",
+            status, body
+        )));
+    }
+
+    let token_data: serde_json::Value = response
+        .json()
+        .await
+        .map_err(|e| OAuthCallbackError::Io(format!("Failed to parse proxy response: {}", e)))?;
+
+    oauth_token_response_from_json(token_data, "access_token")
 }
 
 #[cfg(test)]
 mod tests {
+    use std::collections::HashMap;
+    use std::net::SocketAddr;
+    use std::sync::Arc;
+
+    use axum::extract::{Form, State};
+    use axum::http::HeaderMap;
+    use axum::response::Redirect;
+    use axum::routing::post;
+    use axum::{Json, Router};
+    use serde_json::json;
+    use tokio::net::TcpListener;
+    use tokio::sync::{Mutex, oneshot};
+
     use crate::cli::oauth_defaults::{
         builtin_credentials, callback_host, callback_url, is_loopback_host, landing_html,
     };
     use crate::config::helpers::lock_env;
+    use crate::testing::credentials::{TEST_OAUTH_CLIENT_ID, TEST_OAUTH_CLIENT_SECRET};
+
+    #[derive(Clone, Debug, PartialEq, Eq)]
+    struct RecordedProxyRequest {
+        authorization: Option<String>,
+        form: HashMap<String, String>,
+    }
+
+    #[derive(Clone)]
+    struct MockProxyState {
+        requests: Arc<Mutex<Vec<RecordedProxyRequest>>>,
+        exchange_redirect_target: String,
+        refresh_redirect_target: String,
+    }
+
+    struct MockProxyServer {
+        addr: SocketAddr,
+        requests: Arc<Mutex<Vec<RecordedProxyRequest>>>,
+        shutdown_tx: Option<oneshot::Sender<()>>,
+        server_task: Option<tokio::task::JoinHandle<()>>,
+    }
+
+    impl MockProxyServer {
+        async fn start() -> Self {
+            async fn exchange_handler(
+                State(state): State<MockProxyState>,
+                headers: HeaderMap,
+                Form(form): Form<HashMap<String, String>>,
+            ) -> Json<serde_json::Value> {
+                state.requests.lock().await.push(RecordedProxyRequest {
+                    authorization: headers
+                        .get(axum::http::header::AUTHORIZATION)
+                        .and_then(|value| value.to_str().ok())
+                        .map(str::to_string),
+                    form,
+                });
+                Json(json!({
+                    "access_token": "proxy-access-token",
+                    "refresh_token": "proxy-refresh-token",
+                    "expires_in": 7200
+                }))
+            }
+
+            async fn refresh_handler(
+                State(state): State<MockProxyState>,
+                headers: HeaderMap,
+                Form(form): Form<HashMap<String, String>>,
+            ) -> Json<serde_json::Value> {
+                state.requests.lock().await.push(RecordedProxyRequest {
+                    authorization: headers
+                        .get(axum::http::header::AUTHORIZATION)
+                        .and_then(|value| value.to_str().ok())
+                        .map(str::to_string),
+                    form,
+                });
+                Json(json!({
+                    "access_token": "proxy-access-token",
+                    "refresh_token": "proxy-refresh-token",
+                    "expires_in": 7200
+                }))
+            }
+
+            async fn exchange_redirect_handler(State(state): State<MockProxyState>) -> Redirect {
+                Redirect::temporary(&state.exchange_redirect_target)
+            }
+
+            async fn refresh_redirect_handler(State(state): State<MockProxyState>) -> Redirect {
+                Redirect::temporary(&state.refresh_redirect_target)
+            }
+
+            let requests = Arc::new(Mutex::new(Vec::new()));
+            let listener = TcpListener::bind("127.0.0.1:0")
+                .await
+                .expect("bind mock proxy");
+            let addr = listener.local_addr().expect("read mock proxy addr");
+            let exchange_redirect_target = format!("http://{addr}/oauth/exchange");
+            let refresh_redirect_target = format!("http://{addr}/oauth/refresh");
+            let app = Router::new()
+                .route("/oauth/exchange", post(exchange_handler))
+                .route("/oauth/refresh", post(refresh_handler))
+                .route("/redirect/oauth/exchange", post(exchange_redirect_handler))
+                .route("/redirect/oauth/refresh", post(refresh_redirect_handler))
+                .with_state(MockProxyState {
+                    requests: Arc::clone(&requests),
+                    exchange_redirect_target,
+                    refresh_redirect_target,
+                });
+            let (shutdown_tx, shutdown_rx) = oneshot::channel::<()>();
+            let server_task = tokio::spawn(async move {
+                let _ = axum::serve(listener, app)
+                    .with_graceful_shutdown(async {
+                        let _ = shutdown_rx.await;
+                    })
+                    .await;
+            });
+
+            Self {
+                addr,
+                requests,
+                shutdown_tx: Some(shutdown_tx),
+                server_task: Some(server_task),
+            }
+        }
+
+        fn base_url(&self) -> String {
+            format!("http://{}", self.addr)
+        }
+
+        fn redirecting_base_url(&self) -> String {
+            format!("{}/redirect", self.base_url())
+        }
+
+        async fn requests(&self) -> Vec<RecordedProxyRequest> {
+            self.requests.lock().await.clone()
+        }
+
+        async fn shutdown(mut self) {
+            if let Some(tx) = self.shutdown_tx.take() {
+                let _ = tx.send(());
+            }
+            if let Some(task) = self.server_task.take() {
+                let _ = task.await;
+            }
+        }
+    }
+
+    impl Drop for MockProxyServer {
+        fn drop(&mut self) {
+            if let Some(tx) = self.shutdown_tx.take() {
+                let _ = tx.send(());
+            }
+            if let Some(task) = self.server_task.take() {
+                task.abort();
+            }
+        }
+    }
+
+    #[test]
+    fn test_hosted_proxy_client_secret_suppresses_builtin_secret() {
+        let builtin = builtin_credentials("google_oauth_token").expect("google builtin creds");
+        let client_secret = Some(builtin.client_secret.to_string());
+
+        let result = super::hosted_proxy_client_secret(&client_secret, Some(&builtin), true);
+
+        assert_eq!(result, None);
+    }
+
+    #[test]
+    fn test_hosted_proxy_client_secret_preserves_explicit_secret() {
+        let builtin = builtin_credentials("google_oauth_token").expect("google builtin creds");
+        let client_secret = Some("hosted-server-secret".to_string());
+
+        let result = super::hosted_proxy_client_secret(&client_secret, Some(&builtin), true);
+
+        assert_eq!(result, client_secret);
+    }
+
+    #[tokio::test]
+    async fn test_refresh_token_via_proxy_sends_auth_and_form() {
+        let server = MockProxyServer::start().await;
+
+        let response = super::refresh_token_via_proxy(super::ProxyRefreshTokenRequest {
+            proxy_url: &server.base_url(),
+            gateway_token: "gateway-test-token",
+            token_url: "https://oauth2.googleapis.com/token",
+            client_id: TEST_OAUTH_CLIENT_ID,
+            client_secret: Some(TEST_OAUTH_CLIENT_SECRET),
+            refresh_token: "refresh-token-123",
+            provider: Some("google"),
+        })
+        .await
+        .expect("proxy refresh succeeds");
+
+        assert_eq!(response.access_token, "proxy-access-token");
+        assert_eq!(
+            response.refresh_token.as_deref(),
+            Some("proxy-refresh-token")
+        );
+        assert_eq!(response.expires_in, Some(7200));
+
+        let requests = server.requests().await;
+        assert_eq!(requests.len(), 1);
+        assert_eq!(
+            requests[0].authorization.as_deref(),
+            Some("Bearer gateway-test-token")
+        );
+        assert_eq!(
+            requests[0].form.get("token_url").map(String::as_str),
+            Some("https://oauth2.googleapis.com/token")
+        );
+        assert_eq!(
+            requests[0].form.get("client_id").map(String::as_str),
+            Some(TEST_OAUTH_CLIENT_ID)
+        );
+        assert_eq!(
+            requests[0].form.get("client_secret").map(String::as_str),
+            Some(TEST_OAUTH_CLIENT_SECRET)
+        );
+        assert_eq!(
+            requests[0].form.get("refresh_token").map(String::as_str),
+            Some("refresh-token-123")
+        );
+        assert_eq!(
+            requests[0].form.get("provider").map(String::as_str),
+            Some("google")
+        );
+
+        server.shutdown().await;
+    }
+
+    #[tokio::test]
+    async fn test_exchange_via_proxy_does_not_follow_redirects() {
+        let server = MockProxyServer::start().await;
+
+        let error = match super::exchange_via_proxy(super::ProxyTokenExchangeRequest {
+            proxy_url: &server.redirecting_base_url(),
+            gateway_token: "gateway-test-token",
+            code: "auth-code-123",
+            redirect_uri: "http://localhost:3000/oauth/callback",
+            token_url: "https://oauth2.googleapis.com/token",
+            client_id: TEST_OAUTH_CLIENT_ID,
+            client_secret: Some(TEST_OAUTH_CLIENT_SECRET),
+            access_token_field: "access_token",
+            code_verifier: Some("code-verifier-123"),
+            extra_token_params: &HashMap::new(),
+        })
+        .await
+        {
+            Ok(_) => panic!("redirected proxy exchange should fail"),
+            Err(error) => error,
+        };
+
+        assert!(error.to_string().contains("307"));
+        assert!(server.requests().await.is_empty());
+
+        server.shutdown().await;
+    }
+
+    #[tokio::test]
+    async fn test_refresh_token_via_proxy_does_not_follow_redirects() {
+        let server = MockProxyServer::start().await;
+
+        let error = match super::refresh_token_via_proxy(super::ProxyRefreshTokenRequest {
+            proxy_url: &server.redirecting_base_url(),
+            gateway_token: "gateway-test-token",
+            token_url: "https://oauth2.googleapis.com/token",
+            client_id: TEST_OAUTH_CLIENT_ID,
+            client_secret: Some(TEST_OAUTH_CLIENT_SECRET),
+            refresh_token: "refresh-token-123",
+            provider: Some("google"),
+        })
+        .await
+        {
+            Ok(_) => panic!("redirected proxy refresh should fail"),
+            Err(error) => error,
+        };
+
+        assert!(error.to_string().contains("307"));
+        assert!(server.requests().await.is_empty());
+
+        server.shutdown().await;
+    }
 
     #[test]
     fn test_is_loopback_host() {

src/cli/routines.rs

@@ -10,7 +10,7 @@ use clap::Subcommand;
 use uuid::Uuid;
 
 use crate::agent::routine::{
-    NotifyConfig, Routine, RoutineAction, RoutineGuardrails, Trigger, next_cron_fire,
+    NotifyConfig, Routine, RoutineAction, RoutineGuardrails, RunStatus, Trigger, next_cron_fire,
 };
 use crate::db::Database;
 
@@ -251,15 +251,26 @@ async fn list(
     );
     println!("{}", "-".repeat(130));
 
+    // Fetch last-run status for all routines in a single batch query
+    let routine_ids: Vec<Uuid> = filtered.iter().map(|r| r.id).collect();
+    let last_run_results = db
+        .batch_get_last_run_status(&routine_ids)
+        .await
+        .unwrap_or_default();
+
     for r in &filtered {
-        let status = if r.enabled {
+        let last_run_status = last_run_results.get(&r.id).copied();
-            if r.consecutive_failures > 0 {
+
-                format!("err({})", r.consecutive_failures)
+        let status = if !r.enabled {
-            } else {
-                "active".to_string()
-            }
-        } else {
             "disabled".to_string()
+        } else if last_run_status == Some(RunStatus::Running) {
+            "running".to_string()
+        } else if r.consecutive_failures > 0 {
+            format!("err({})", r.consecutive_failures)
+        } else if last_run_status == Some(RunStatus::Attention) {
+            "attention".to_string()
+        } else {
+            "active".to_string()
         };
 
         let next_fire = r
@@ -340,8 +351,8 @@ async fn create(
             prompt: prompt.to_string(),
             context_paths: Vec::new(),
             max_tokens: 4096,
-            use_tools: false,
+            use_tools: true,
-            max_tool_rounds: 0,
+            max_tool_rounds: 3,
         },
         guardrails: RoutineGuardrails {
             cooldown: std::time::Duration::from_secs(cooldown_secs),
@@ -685,6 +696,7 @@ fn truncate(s: &str, max_chars: usize) -> String {
 #[cfg(test)]
 mod tests {
     use super::*;
+    use crate::agent::routine::RoutineAction;
 
     #[test]
     fn format_relative_future() {
@@ -743,4 +755,48 @@ mod tests {
         assert!(notify.on_failure); // safety: test-only assertion
         assert!(!notify.on_success); // safety: test-only assertion
     }
+
+    #[cfg(feature = "libsql")]
+    #[tokio::test]
+    async fn cli_create_defaults_lightweight_routines_to_tools_enabled() {
+        let harness = crate::testing::TestHarnessBuilder::new().build().await;
+        let db = harness.db.clone();
+
+        run_routines_command(
+            RoutinesCommand::Create {
+                name: "cli-digest".to_string(),
+                schedule: "0 0 9 * * *".to_string(),
+                prompt: "Prepare the morning digest.".to_string(),
+                description: "CLI created routine".to_string(),
+                timezone: Some("UTC".to_string()),
+                cooldown: 300,
+                notify_channel: None,
+            },
+            db.clone(),
+            "user1",
+        )
+        .await
+        .expect("create routine");
+
+        let routine = db
+            .get_routine_by_name("user1", "cli-digest")
+            .await
+            .expect("get routine by name")
+            .expect("cli-digest should exist");
+
+        match routine.action {
+            RoutineAction::Lightweight {
+                use_tools,
+                max_tool_rounds,
+                ..
+            } => {
+                assert!(
+                    use_tools,
+                    "CLI-created lightweight routines should default to tools"
+                );
+                assert_eq!(max_tool_rounds, 3);
+            }
+            other => panic!("expected lightweight action, got {other:?}"),
+        }
+    }
 }

src/cli/snapshots/ironclaw__cli__tests__help_output.snap

@@ -20,6 +20,7 @@ Commands:
   service     Manage OS service
   skills      Manage skills
   hooks       Manage lifecycle hooks
+  models      Manage LLM providers and models
   doctor      Run diagnostics
   logs        View and manage gateway logs
   status      Show system status

src/cli/snapshots/ironclaw__cli__tests__help_output_without_import.snap

@@ -20,6 +20,7 @@ Commands:
   service     Manage OS service
   skills      Manage skills
   hooks       Manage lifecycle hooks
+  models      Manage LLM providers and models
   doctor      Run diagnostics
   logs        View and manage gateway logs
   status      Show system status

src/cli/snapshots/ironclaw__cli__tests__long_help_output.snap

@@ -23,6 +23,7 @@ Commands:
   service     Manage OS service
   skills      Manage skills
   hooks       Manage lifecycle hooks
+  models      Manage LLM providers and models
   doctor      Run diagnostics
   logs        View and manage gateway logs
   status      Show system status

src/cli/snapshots/ironclaw__cli__tests__long_help_output_without_import.snap

@@ -23,6 +23,7 @@ Commands:
   service     Manage OS service
   skills      Manage skills
   hooks       Manage lifecycle hooks
+  models      Manage LLM providers and models
   doctor      Run diagnostics
   logs        View and manage gateway logs
   status      Show system status

src/cli/tool.rs

@@ -2,6 +2,7 @@
 //!
 //! Commands for installing, listing, removing, and authenticating WASM tools.
 
+use std::collections::{HashMap, HashSet};
 use std::io::Write;
 use std::path::{Path, PathBuf};
 use std::sync::Arc;
@@ -79,6 +80,10 @@ pub enum ToolCommand {
         /// Directory to look for tool (default: ~/.ironclaw/tools/)
         #[arg(short, long)]
         dir: Option<PathBuf>,
+
+        /// User ID for checking credential status (default: "default")
+        #[arg(short, long, default_value = "default")]
+        user: String,
     },
 
     /// Configure authentication for a tool
@@ -124,7 +129,11 @@ pub async fn run_tool_command(cmd: ToolCommand) -> anyhow::Result<()> {
         } => install_tool(path, name, capabilities, target, release, skip_build, force).await,
         ToolCommand::List { dir, verbose } => list_tools(dir, verbose).await,
         ToolCommand::Remove { name, dir } => remove_tool(name, dir).await,
-        ToolCommand::Info { name_or_path, dir } => show_tool_info(name_or_path, dir).await,
+        ToolCommand::Info {
+            name_or_path,
+            dir,
+            user,
+        } => show_tool_info(name_or_path, dir, user).await,
         ToolCommand::Auth { name, dir, user } => auth_tool(name, dir, user).await,
         ToolCommand::Setup { name, dir, user } => setup_tool(name, dir, user).await,
     }
@@ -388,7 +397,11 @@ async fn remove_tool(name: String, dir: Option<PathBuf>) -> anyhow::Result<()> {
 }
 
 /// Show information about a tool.
-async fn show_tool_info(name_or_path: String, dir: Option<PathBuf>) -> anyhow::Result<()> {
+async fn show_tool_info(
+    name_or_path: String,
+    dir: Option<PathBuf>,
+    user_id: String,
+) -> anyhow::Result<()> {
     let wasm_path = if name_or_path.ends_with(".wasm") {
         PathBuf::from(&name_or_path)
     } else {
@@ -423,7 +436,37 @@ async fn show_tool_info(name_or_path: String, dir: Option<PathBuf>) -> anyhow::R
         println!("\nCapabilities ({}):", caps_path.display());
         let content = fs::read_to_string(&caps_path).await?;
         match CapabilitiesFile::from_json(&content) {
-            Ok(caps) => print_capabilities_detail(&caps),
+            Ok(caps) => {
+                // Lazily init secrets store only when auth secrets need checking.
+                let has_auth = caps.auth.is_some()
+                    || caps
+                        .setup
+                        .as_ref()
+                        .is_some_and(|s| !s.required_secrets.is_empty())
+                    || caps
+                        .http
+                        .as_ref()
+                        .is_some_and(|h| !h.credentials.is_empty());
+                let secrets_store = if has_auth {
+                    match init_secrets_store().await {
+                        Ok(store) => Some(store),
+                        Err(e) => {
+                            eprintln!("  Warning: could not init secrets store: {}", e);
+                            None
+                        }
+                    }
+                } else {
+                    None
+                };
+                print_capabilities_detail(
+                    &caps,
+                    secrets_store
+                        .as_ref()
+                        .map(|s| s.as_ref() as &(dyn SecretsStore + Send + Sync)),
+                    &user_id,
+                )
+                .await;
+            }
             Err(e) => println!("  Error parsing: {}", e),
         }
     } else {
@@ -476,8 +519,89 @@ fn print_capabilities_summary(caps: &CapabilitiesFile) {
     }
 }
 
+/// Per-secret info collected from all auth-related capability sections.
+struct AuthSecretInfo {
+    secret_name: String,
+    /// Human-readable label (from auth.display_name or setup prompt).
+    description: Option<String>,
+    /// Injection location (from http.credentials).
+    location: Option<String>,
+}
+
+/// Collected auth secrets and the set of secret names they cover.
+struct CollectedAuthSecrets {
+    secrets: Vec<AuthSecretInfo>,
+    /// Secret names present in `secrets`, for filtering the Secrets capability section.
+    seen_names: HashSet<String>,
+}
+
+/// Collect and deduplicate auth secrets from all auth-related capability sections.
+///
+/// Priority for the description label: auth.display_name > setup.required_secrets.prompt.
+/// Injection location is merged from http.credentials.
+fn collect_auth_secrets(caps: &CapabilitiesFile) -> CollectedAuthSecrets {
+    let mut secrets: Vec<AuthSecretInfo> = Vec::new();
+    let mut seen: HashMap<String, usize> = HashMap::new();
+
+    // auth.display_name is the best label — seed first.
+    if let Some(ref auth) = caps.auth {
+        let index = secrets.len();
+        seen.insert(auth.secret_name.clone(), index);
+        secrets.push(AuthSecretInfo {
+            secret_name: auth.secret_name.clone(),
+            description: auth.display_name.clone(),
+            location: None,
+        });
+    }
+
+    // setup.required_secrets.prompt is second-best label.
+    if let Some(ref setup) = caps.setup {
+        for secret in &setup.required_secrets {
+            if !seen.contains_key(&secret.name) {
+                let index = secrets.len();
+                seen.insert(secret.name.clone(), index);
+                secrets.push(AuthSecretInfo {
+                    secret_name: secret.name.clone(),
+                    description: Some(secret.prompt.clone()),
+                    location: None,
+                });
+            }
+        }
+    }
+
+    // Merge injection location from http.credentials.
+    if let Some(ref http) = caps.http {
+        for cred in http.credentials.values() {
+            let loc = format!("{:?}", cred.location);
+            if let Some(&index) = seen.get(&cred.secret_name) {
+                secrets[index].location = Some(loc);
+            } else {
+                let index = secrets.len();
+                seen.insert(cred.secret_name.clone(), index);
+                secrets.push(AuthSecretInfo {
+                    secret_name: cred.secret_name.clone(),
+                    description: None,
+                    location: Some(loc),
+                });
+            }
+        }
+    }
+
+    let seen_names = seen.into_keys().collect();
+    CollectedAuthSecrets {
+        secrets,
+        seen_names,
+    }
+}
+
 /// Print detailed capabilities.
-fn print_capabilities_detail(caps: &CapabilitiesFile) {
+async fn print_capabilities_detail(
+    caps: &CapabilitiesFile,
+    secrets_store: Option<&(dyn SecretsStore + Send + Sync)>,
+    user_id: &str,
+) {
+    let mut collected = collect_auth_secrets(caps);
+
     if let Some(ref http) = caps.http {
         println!("  HTTP:");
         for endpoint in &http.allowlist {
@@ -490,13 +614,6 @@ fn print_capabilities_detail(caps: &CapabilitiesFile) {
             println!("    {} {} {}", methods, endpoint.host, path);
         }
 
-        if !http.credentials.is_empty() {
-            println!("  Credentials:");
-            for (key, cred) in &http.credentials {
-                println!("    {}: {} -> {:?}", key, cred.secret_name, cred.location);
-            }
-        }
-
         if let Some(ref rate) = http.rate_limit {
             println!(
                 "  Rate limit: {}/min, {}/hour",
@@ -505,12 +622,24 @@ fn print_capabilities_detail(caps: &CapabilitiesFile) {
         }
     }
 
+    // Filter secrets already covered by the auth section (always rendered when non-empty).
     if let Some(ref secrets) = caps.secrets
         && !secrets.allowed_names.is_empty()
     {
-        println!("  Secrets (existence check only):");
+        let extra: Vec<_> = if collected.secrets.is_empty() {
-        for name in &secrets.allowed_names {
+            secrets.allowed_names.iter().collect()
-            println!("    {}", name);
+        } else {
+            secrets
+                .allowed_names
+                .iter()
+                .filter(|name| !collected.seen_names.contains(name.as_str()))
+                .collect()
+        };
+        if !extra.is_empty() {
+            println!("  Secrets (existence check only):");
+            for name in extra {
+                println!("    {}", name);
+            }
         }
     }
 
@@ -531,6 +660,38 @@ fn print_capabilities_detail(caps: &CapabilitiesFile) {
             println!("    {}", prefix);
         }
     }
+
+    // Consolidated auth status — sorted by secret name for deterministic output.
+    if !collected.secrets.is_empty() {
+        collected
+            .secrets
+            .sort_by(|a, b| a.secret_name.cmp(&b.secret_name));
+        println!("  Auth:");
+        for info in &collected.secrets {
+            let (icon, label) = match secrets_store {
+                Some(store) => match store.exists(user_id, &info.secret_name).await {
+                    Ok(true) => ("\u{2713}", "configured"),
+                    Ok(false) => ("\u{2717}", "missing"),
+                    Err(e) => {
+                        eprintln!(
+                            "  Warning: failed to check secret `{}`: {}",
+                            info.secret_name, e
+                        );
+                        ("?", "unknown")
+                    }
+                },
+                None => ("?", "unknown"),
+            };
+            let mut parts = info.secret_name.clone();
+            if let Some(ref desc) = info.description {
+                parts = format!("{} ({})", parts, desc);
+            }
+            if let Some(ref loc) = info.location {
+                parts = format!("{} -> {}", parts, loc);
+            }
+            println!("    {}  {} {}", parts, icon, label);
+        }
+    }
 }
 
 /// Validate a tool name to prevent path traversal.
@@ -677,8 +838,7 @@ async fn combine_provider_scopes(
     secret_name: &str,
     base_oauth: &crate::tools::wasm::OAuthConfigSchema,
 ) -> crate::tools::wasm::OAuthConfigSchema {
-    let mut all_scopes: std::collections::HashSet<String> =
+    let mut all_scopes: HashSet<String> = base_oauth.scopes.iter().cloned().collect();
-        base_oauth.scopes.iter().cloned().collect();
 
     if let Ok(mut entries) = tokio::fs::read_dir(tools_dir).await {
         while let Ok(Some(entry)) = entries.next_entry().await {
@@ -1127,6 +1287,8 @@ async fn setup_tool(name: String, dir: Option<PathBuf>, user_id: String) -> anyh
 #[cfg(test)]
 mod tests {
     use super::*;
+    use crate::secrets::{CreateSecretParams, SecretsStore};
+    use crate::testing::credentials::test_secrets_store;
 
     #[test]
     fn test_format_size() {
@@ -1143,4 +1305,96 @@ mod tests {
         assert!(dir.to_string_lossy().contains(".ironclaw"));
         assert!(dir.to_string_lossy().contains("tools"));
     }
+
+    /// Verify that auth secrets are deduplicated across auth, setup, and http.credentials,
+    /// and that credential status is checked against the secrets store.
+    #[tokio::test]
+    async fn test_auth_secret_dedup_and_status() {
+        let caps = CapabilitiesFile::from_json(
+            r#"{
+                "auth": {
+                    "secret_name": "gh_token",
+                    "display_name": "GitHub"
+                },
+                "setup": {
+                    "required_secrets": [
+                        { "name": "gh_token", "prompt": "GitHub PAT" },
+                        { "name": "extra_key", "prompt": "Extra API Key" }
+                    ]
+                },
+                "http": {
+                    "allowlist": [{ "host": "api.github.com" }],
+                    "credentials": {
+                        "github": {
+                            "secret_name": "gh_token",
+                            "location": { "type": "bearer" },
+                            "host_patterns": ["api.github.com"]
+                        }
+                    }
+                },
+                "secrets": {
+                    "allowed_names": ["gh_token", "gh_*"]
+                }
+            }"#,
+        )
+        .unwrap();
+
+        let collected = collect_auth_secrets(&caps);
+
+        // gh_token should appear once (from auth), with location merged from credentials.
+        // extra_key should appear once (from setup).
+        assert_eq!(collected.secrets.len(), 2);
+        let gh = collected
+            .secrets
+            .iter()
+            .find(|s| s.secret_name == "gh_token")
+            .unwrap();
+        assert_eq!(gh.description.as_deref(), Some("GitHub"));
+        assert!(
+            gh.location.is_some(),
+            "location should be merged from http.credentials"
+        );
+
+        let extra = collected
+            .secrets
+            .iter()
+            .find(|s| s.secret_name == "extra_key")
+            .unwrap();
+        assert_eq!(extra.description.as_deref(), Some("Extra API Key"));
+        assert!(extra.location.is_none());
+
+        // Secrets section should filter gh_token (in seen_names) but keep gh_* (wildcard).
+        let secrets = caps.secrets.as_ref().unwrap();
+        let extra_secrets: Vec<_> = secrets
+            .allowed_names
+            .iter()
+            .filter(|name| !collected.seen_names.contains(name.as_str()))
+            .collect();
+        assert_eq!(extra_secrets, vec!["gh_*"]);
+
+        // Verify store check: missing secret -> exists returns false.
+        let store = test_secrets_store();
+        assert!(!store.exists("default", "gh_token").await.unwrap());
+
+        // Store gh_token and verify it's found.
+        store
+            .create(
+                "default",
+                CreateSecretParams::new("gh_token", "ghp_test123"),
+            )
+            .await
+            .unwrap();
+        assert!(store.exists("default", "gh_token").await.unwrap());
+        // extra_key still missing.
+        assert!(!store.exists("default", "extra_key").await.unwrap());
+    }
+
+    /// No auth sections → collect_auth_secrets returns empty.
+    #[test]
+    fn test_collect_auth_secrets_empty_caps() {
+        let caps = CapabilitiesFile::default();
+        let collected = collect_auth_secrets(&caps);
+        assert!(collected.secrets.is_empty());
+        assert!(collected.seen_names.is_empty());
+    }
 }

src/config/channels.rs

@@ -2,6 +2,7 @@ use std::collections::HashMap;
 use std::path::PathBuf;
 
 use secrecy::SecretString;
+use serde::Deserialize;
 
 use crate::bootstrap::ironclaw_base_dir;
 use crate::config::helpers::{optional_env, parse_bool_env, parse_optional_env};
@@ -45,6 +46,26 @@ pub struct GatewayConfig {
     /// Bearer token for authentication. Random hex generated at startup if unset.
     pub auth_token: Option<String>,
     pub user_id: String,
+    /// Additional user scopes for workspace reads.
+    ///
+    /// When set, the workspace will be able to read (search, read, list) from
+    /// these additional user scopes while writes remain isolated to `user_id`.
+    /// Parsed from `WORKSPACE_READ_SCOPES` (comma-separated).
+    pub workspace_read_scopes: Vec<String>,
+    /// Memory layer definitions (JSON in env var, or from external config).
+    pub memory_layers: Vec<crate::workspace::layer::MemoryLayer>,
+    /// Multi-user token map. When set, each token maps to a user identity.
+    /// Parsed from `GATEWAY_USER_TOKENS` (JSON string). When absent, falls back
+    /// to single-user mode via `auth_token` + `user_id`.
+    pub user_tokens: Option<HashMap<String, UserTokenConfig>>,
+}
+
+/// Per-user token configuration for multi-user mode.
+#[derive(Debug, Clone, Deserialize)]
+pub struct UserTokenConfig {
+    pub user_id: String,
+    #[serde(default)]
+    pub workspace_read_scopes: Vec<String>,
 }
 
 /// Signal channel configuration (signal-cli daemon HTTP/JSON-RPC).
@@ -115,6 +136,118 @@ impl ChannelsConfig {
                 .or_else(|| cs.gateway_user_id.clone())
                 .unwrap_or_else(|| owner_id.to_string());
 
+            let memory_layers: Vec<crate::workspace::layer::MemoryLayer> =
+                match optional_env("MEMORY_LAYERS")? {
+                    Some(json_str) => {
+                        serde_json::from_str(&json_str).map_err(|e| ConfigError::InvalidValue {
+                            key: "MEMORY_LAYERS".to_string(),
+                            message: format!("must be valid JSON array of layer objects: {e}"),
+                        })?
+                    }
+                    None => crate::workspace::layer::MemoryLayer::default_for_user(&user_id),
+                };
+
+            // Validate layer names and scopes
+            for layer in &memory_layers {
+                if layer.name.trim().is_empty() {
+                    return Err(ConfigError::InvalidValue {
+                        key: "MEMORY_LAYERS".to_string(),
+                        message: "layer name must not be empty".to_string(),
+                    });
+                }
+                if layer.name.len() > 64 {
+                    return Err(ConfigError::InvalidValue {
+                        key: "MEMORY_LAYERS".to_string(),
+                        message: format!("layer name '{}' exceeds 64 characters", layer.name),
+                    });
+                }
+                if !layer
+                    .name
+                    .chars()
+                    .all(|c| c.is_ascii_alphanumeric() || c == '_' || c == '-')
+                {
+                    return Err(ConfigError::InvalidValue {
+                        key: "MEMORY_LAYERS".to_string(),
+                        message: format!(
+                            "layer name '{}' contains invalid characters \
+                             (allowed: a-z, A-Z, 0-9, _, -)",
+                            layer.name
+                        ),
+                    });
+                }
+                if layer.scope.trim().is_empty() {
+                    return Err(ConfigError::InvalidValue {
+                        key: "MEMORY_LAYERS".to_string(),
+                        message: format!("layer '{}' has an empty scope", layer.name),
+                    });
+                }
+            }
+
+            // Check for duplicate layer names
+            {
+                let mut seen = std::collections::HashSet::new();
+                for layer in &memory_layers {
+                    if !seen.insert(&layer.name) {
+                        return Err(ConfigError::InvalidValue {
+                            key: "MEMORY_LAYERS".to_string(),
+                            message: format!("duplicate layer name '{}'", layer.name),
+                        });
+                    }
+                }
+            }
+
+            let user_tokens: Option<HashMap<String, UserTokenConfig>> =
+                match optional_env("GATEWAY_USER_TOKENS")? {
+                    Some(json_str) => {
+                        let tokens: HashMap<String, UserTokenConfig> = serde_json::from_str(
+                            &json_str,
+                        )
+                        .map_err(|e| ConfigError::InvalidValue {
+                            key: "GATEWAY_USER_TOKENS".to_string(),
+                            message: format!(
+                                "must be valid JSON object mapping tokens to user configs: {e}"
+                            ),
+                        })?;
+                        if tokens.is_empty() {
+                            return Err(ConfigError::InvalidValue {
+                            key: "GATEWAY_USER_TOKENS".to_string(),
+                            message:
+                                "token map is empty — remove the variable to use single-user mode"
+                                    .to_string(),
+                        });
+                        }
+                        for (tok, cfg) in &tokens {
+                            if cfg.user_id.trim().is_empty() {
+                                return Err(ConfigError::InvalidValue {
+                                    key: "GATEWAY_USER_TOKENS".to_string(),
+                                    message: format!(
+                                        "token '{}...' has an empty user_id",
+                                        &tok[..tok.len().min(8)]
+                                    ),
+                                });
+                            }
+                        }
+                        Some(tokens)
+                    }
+                    None => None,
+                };
+            let workspace_read_scopes: Vec<String> = optional_env("WORKSPACE_READ_SCOPES")?
+                .map(|s| {
+                    s.split(',')
+                        .map(|s| s.trim().to_string())
+                        .filter(|s| !s.is_empty())
+                        .collect()
+                })
+                .unwrap_or_default();
+
+            for scope in &workspace_read_scopes {
+                if scope.len() > 128 {
+                    return Err(ConfigError::InvalidValue {
+                        key: "WORKSPACE_READ_SCOPES".to_string(),
+                        message: format!("scope '{}...' exceeds 128 characters", &scope[..32]),
+                    });
+                }
+            }
             Some(GatewayConfig {
                 host: optional_env("GATEWAY_HOST")?
                     .or_else(|| cs.gateway_host.clone())
@@ -126,6 +259,9 @@ impl ChannelsConfig {
                 auth_token: optional_env("GATEWAY_AUTH_TOKEN")?
                     .or_else(|| cs.gateway_auth_token.clone()),
                 user_id,
+                workspace_read_scopes,
+                memory_layers,
+                user_tokens,
             })
         } else {
             None
@@ -281,6 +417,9 @@ mod tests {
             port: 3000,
             auth_token: Some("tok-abc".to_string()),
             user_id: "default".to_string(),
+            workspace_read_scopes: vec![],
+            memory_layers: vec![],
+            user_tokens: None,
         };
         assert_eq!(cfg.host, "127.0.0.1");
         assert_eq!(cfg.port, 3000);
@@ -295,6 +434,9 @@ mod tests {
             port: 3001,
             auth_token: None,
             user_id: "anon".to_string(),
+            workspace_read_scopes: vec![],
+            memory_layers: vec![],
+            user_tokens: None,
         };
         assert!(cfg.auth_token.is_none());
     }

src/config/llm.rs

@@ -406,7 +406,7 @@ impl LlmConfig {
         // Resolve extra headers
         let extra_headers = if let Some(env_var) = extra_headers_env {
             optional_env(env_var)?
-                .map(|val| parse_extra_headers(&val))
+                .map(|val| parse_extra_headers_with_key(&val, env_var))
                 .transpose()?
                 .unwrap_or_default()
         } else {
@@ -475,7 +475,10 @@ impl LlmConfig {
 ///
 /// Format: `Key1:Value1,Key2:Value2` (colon-separated, not `=`, because
 /// header values often contain `=`).
-fn parse_extra_headers(val: &str) -> Result<Vec<(String, String)>, ConfigError> {
+fn parse_extra_headers_with_key(
+    val: &str,
+    env_var_name: &str,
+) -> Result<Vec<(String, String)>, ConfigError> {
     if val.trim().is_empty() {
         return Ok(Vec::new());
     }
@@ -488,14 +491,14 @@ fn parse_extra_headers(val: &str) -> Result<Vec<(String, String)>, ConfigError>
         }
         let Some((key, value)) = pair.split_once(':') else {
             return Err(ConfigError::InvalidValue {
-                key: "LLM_EXTRA_HEADERS".to_string(),
+                key: env_var_name.to_string(),
                 message: format!("malformed header entry '{}', expected Key:Value", pair),
             });
         };
         let key = key.trim();
         if key.is_empty() {
             return Err(ConfigError::InvalidValue {
-                key: "LLM_EXTRA_HEADERS".to_string(),
+                key: env_var_name.to_string(),
                 message: format!("empty header name in entry '{}'", pair),
             });
         }
@@ -536,6 +539,11 @@ mod tests {
     use crate::settings::Settings;
     use crate::testing::credentials::*;
 
+    /// Convenience wrapper for tests — uses "TEST_HEADERS" as the env var name.
+    fn parse_extra_headers(val: &str) -> Result<Vec<(String, String)>, ConfigError> {
+        parse_extra_headers_with_key(val, "TEST_HEADERS")
+    }
+
     /// Clear all openai-compatible-related env vars.
     fn clear_openai_compatible_env() {
         // SAFETY: Only called under ENV_MUTEX in tests.

src/config/mod.rs

@@ -312,13 +312,11 @@ impl Config {
         let tunnel = TunnelConfig::resolve(settings)?;
         let channels = ChannelsConfig::resolve(settings, &owner_id)?;
 
-        // Resolve workspace config using the gateway user_id for default layers.
+        // Resolve the startup workspace against the durable owner scope. The
-        let workspace_user_id = channels
+        // gateway may expose a distinct sender identity, but the base runtime
-            .gateway
+        // workspace stays owner-scoped and per-user gateway workspaces are
-            .as_ref()
+        // handled separately by WorkspacePool.
-            .map(|gw| gw.user_id.as_str())
+        let workspace = WorkspaceConfig::resolve(&owner_id)?;
-            .unwrap_or("default");
-        let workspace = WorkspaceConfig::resolve(workspace_user_id)?;
 
         Ok(Self {
             owner_id: owner_id.clone(),

src/db/libsql/jobs.rs

@@ -230,6 +230,49 @@ impl JobStore for LibSqlBackend {
         Ok(jobs)
     }
 
+    async fn list_agent_jobs_for_user(
+        &self,
+        user_id: &str,
+    ) -> Result<Vec<AgentJobRecord>, DatabaseError> {
+        let conn = self.connect().await?;
+        let mut rows = conn
+            .query(
+                r#"
+                SELECT id, title, status, user_id, failure_reason,
+                       created_at, started_at, completed_at
+                FROM agent_jobs WHERE source = 'direct' AND user_id = ?1
+                ORDER BY created_at DESC
+                "#,
+                params![user_id],
+            )
+            .await
+            .map_err(|e| DatabaseError::Query(e.to_string()))?;
+
+        let mut jobs = Vec::new();
+        while let Some(row) = rows
+            .next()
+            .await
+            .map_err(|e| DatabaseError::Query(e.to_string()))?
+        {
+            let id_str = get_text(&row, 0);
+            let Ok(id) = id_str.parse() else {
+                tracing::warn!("Skipping agent job with invalid UUID: {}", id_str);
+                continue;
+            };
+            jobs.push(AgentJobRecord {
+                id,
+                title: get_text(&row, 1),
+                status: get_text(&row, 2),
+                user_id: get_text(&row, 3),
+                failure_reason: get_opt_text(&row, 4),
+                created_at: get_ts(&row, 5),
+                started_at: get_opt_ts(&row, 6),
+                completed_at: get_opt_ts(&row, 7),
+            });
+        }
+        Ok(jobs)
+    }
+
     async fn get_agent_job_failure_reason(
         &self,
         id: Uuid,
@@ -277,6 +320,32 @@ impl JobStore for LibSqlBackend {
         Ok(summary)
     }
 
+    async fn agent_job_summary_for_user(
+        &self,
+        user_id: &str,
+    ) -> Result<AgentJobSummary, DatabaseError> {
+        let conn = self.connect().await?;
+        let mut rows = conn
+            .query(
+                "SELECT status, COUNT(*) as cnt FROM agent_jobs WHERE source = 'direct' AND user_id = ?1 GROUP BY status",
+                params![user_id],
+            )
+            .await
+            .map_err(|e| DatabaseError::Query(e.to_string()))?;
+
+        let mut summary = AgentJobSummary::default();
+        while let Some(row) = rows
+            .next()
+            .await
+            .map_err(|e| DatabaseError::Query(e.to_string()))?
+        {
+            let status = get_text(&row, 0);
+            let count = get_i64(&row, 1) as usize;
+            summary.add_count(&status, count);
+        }
+        Ok(summary)
+    }
+
     async fn save_action(&self, job_id: Uuid, action: &ActionRecord) -> Result<(), DatabaseError> {
         let conn = self.connect().await?;
         let duration_ms = action.duration.as_millis() as i64;

src/db/libsql/routines.rs

@@ -462,6 +462,56 @@ impl RoutineStore for LibSqlBackend {
         Ok(counts)
     }
 
+    async fn batch_get_last_run_status(
+        &self,
+        routine_ids: &[Uuid],
+    ) -> Result<HashMap<Uuid, RunStatus>, DatabaseError> {
+        if routine_ids.is_empty() {
+            return Ok(HashMap::new());
+        }
+
+        let conn = self.connect().await?;
+
+        // SQLite doesn't support ANY($1), so we query all latest runs and filter in memory.
+        // Uses a subquery to pick only the most recent run per routine.
+        let mut rows = conn
+            .query(
+                "SELECT routine_id, status FROM routine_runs r1
+                 WHERE started_at = (
+                     SELECT MAX(started_at) FROM routine_runs r2
+                     WHERE r2.routine_id = r1.routine_id
+                 )
+                 GROUP BY routine_id",
+                params![],
+            )
+            .await
+            .map_err(|e| {
+                DatabaseError::Query(format!("Failed to batch get last run status: {}", e))
+            })?;
+
+        let routine_id_set: HashSet<Uuid> = routine_ids.iter().copied().collect();
+        let mut statuses = HashMap::new();
+
+        while let Some(row) = rows
+            .next()
+            .await
+            .map_err(|e| DatabaseError::Query(e.to_string()))?
+        {
+            let id_str: String = get_text(&row, 0);
+            let id = Uuid::parse_str(&id_str)
+                .map_err(|e| DatabaseError::Query(format!("Invalid routine UUID: {}", e)))?;
+
+            if routine_id_set.contains(&id) {
+                let status_str: String = get_text(&row, 1);
+                if let std::result::Result::Ok(status) = status_str.parse::<RunStatus>() {
+                    statuses.insert(id, status);
+                }
+            }
+        }
+
+        Ok(statuses)
+    }
+
     async fn link_routine_run_to_job(
         &self,
         run_id: Uuid,

src/db/mod.rs

@@ -409,7 +409,15 @@ pub trait JobStore: Send + Sync {
     async fn mark_job_stuck(&self, id: Uuid) -> Result<(), DatabaseError>;
     async fn get_stuck_jobs(&self) -> Result<Vec<Uuid>, DatabaseError>;
     async fn list_agent_jobs(&self) -> Result<Vec<AgentJobRecord>, DatabaseError>;
+    async fn list_agent_jobs_for_user(
+        &self,
+        user_id: &str,
+    ) -> Result<Vec<AgentJobRecord>, DatabaseError>;
     async fn agent_job_summary(&self) -> Result<AgentJobSummary, DatabaseError>;
+    async fn agent_job_summary_for_user(
+        &self,
+        user_id: &str,
+    ) -> Result<AgentJobSummary, DatabaseError>;
     /// Get the failure reason for a single agent job (O(1) lookup).
     async fn get_agent_job_failure_reason(&self, id: Uuid)
     -> Result<Option<String>, DatabaseError>;
@@ -520,6 +528,15 @@ pub trait RoutineStore: Send + Sync {
         &self,
         routine_ids: &[Uuid],
     ) -> Result<HashMap<Uuid, i64>, DatabaseError>;
+
+    /// Fetch the last run status for multiple routines in a single query.
+    /// Returns a map from routine_id to its most recent RunStatus.
+    /// Routines with no runs are omitted from the result.
+    async fn batch_get_last_run_status(
+        &self,
+        routine_ids: &[Uuid],
+    ) -> Result<HashMap<Uuid, RunStatus>, DatabaseError>;
+
     async fn link_routine_run_to_job(
         &self,
         run_id: Uuid,

src/db/postgres.rs

@@ -249,10 +249,24 @@ impl JobStore for PgBackend {
         self.store.list_agent_jobs().await
     }
 
+    async fn list_agent_jobs_for_user(
+        &self,
+        user_id: &str,
+    ) -> Result<Vec<AgentJobRecord>, DatabaseError> {
+        self.store.list_agent_jobs_for_user(user_id).await
+    }
+
     async fn agent_job_summary(&self) -> Result<AgentJobSummary, DatabaseError> {
         self.store.agent_job_summary().await
     }
 
+    async fn agent_job_summary_for_user(
+        &self,
+        user_id: &str,
+    ) -> Result<AgentJobSummary, DatabaseError> {
+        self.store.agent_job_summary_for_user(user_id).await
+    }
+
     async fn get_agent_job_failure_reason(
         &self,
         id: Uuid,
@@ -496,6 +510,14 @@ impl RoutineStore for PgBackend {
             .await
     }
 
+    async fn batch_get_last_run_status(
+        &self,
+        routine_ids: &[Uuid],
+    ) -> Result<std::collections::HashMap<Uuid, crate::agent::routine::RunStatus>, DatabaseError>
+    {
+        self.store.batch_get_last_run_status(routine_ids).await
+    }
+
     async fn link_routine_run_to_job(
         &self,
         run_id: Uuid,

src/history/store.rs

@@ -842,6 +842,38 @@ impl Store {
             .collect())
     }
 
+    pub async fn list_agent_jobs_for_user(
+        &self,
+        user_id: &str,
+    ) -> Result<Vec<AgentJobRecord>, DatabaseError> {
+        let conn = self.conn().await?;
+        let rows = conn
+            .query(
+                r#"
+                SELECT id, title, status, user_id, failure_reason,
+                       created_at, started_at, completed_at
+                FROM agent_jobs WHERE source = 'direct' AND user_id = $1
+                ORDER BY created_at DESC
+                "#,
+                &[&user_id],
+            )
+            .await?;
+
+        Ok(rows
+            .iter()
+            .map(|r| AgentJobRecord {
+                id: r.get("id"),
+                title: r.get("title"),
+                status: r.get("status"),
+                user_id: r.get::<_, Option<String>>("user_id").unwrap_or_default(),
+                created_at: r.get("created_at"),
+                started_at: r.get("started_at"),
+                completed_at: r.get("completed_at"),
+                failure_reason: r.get("failure_reason"),
+            })
+            .collect())
+    }
+
     /// Get the failure reason for a single agent job.
     pub async fn get_agent_job_failure_reason(
         &self,
@@ -875,6 +907,27 @@ impl Store {
         }
         Ok(summary)
     }
+
+    pub async fn agent_job_summary_for_user(
+        &self,
+        user_id: &str,
+    ) -> Result<AgentJobSummary, DatabaseError> {
+        let conn = self.conn().await?;
+        let rows = conn
+            .query(
+                "SELECT status, COUNT(*) as cnt FROM agent_jobs WHERE source = 'direct' AND user_id = $1 GROUP BY status",
+                &[&user_id],
+            )
+            .await?;
+
+        let mut summary = AgentJobSummary::default();
+        for row in &rows {
+            let status: String = row.get("status");
+            let count: i64 = row.get("cnt");
+            summary.add_count(&status, count as usize);
+        }
+        Ok(summary)
+    }
 }
 
 // ==================== Job Events ====================
@@ -1350,6 +1403,40 @@ impl Store {
         Ok(counts)
     }
 
+    /// Batch-load the most recent run status for multiple routines in a single query.
+    /// Uses a window function to pick only the latest run per routine.
+    #[cfg(feature = "postgres")]
+    pub async fn batch_get_last_run_status(
+        &self,
+        routine_ids: &[Uuid],
+    ) -> Result<HashMap<Uuid, RunStatus>, DatabaseError> {
+        if routine_ids.is_empty() {
+            return Ok(HashMap::new());
+        }
+
+        let conn = self.conn().await?;
+        let rows = conn
+            .query(
+                "SELECT DISTINCT ON (routine_id) routine_id, status
+                 FROM routine_runs
+                 WHERE routine_id = ANY($1)
+                 ORDER BY routine_id, started_at DESC",
+                &[&routine_ids],
+            )
+            .await?;
+
+        let mut statuses = HashMap::new();
+        for row in rows {
+            let id: Uuid = row.get("routine_id");
+            let status_str: String = row.get("status");
+            if let std::result::Result::Ok(status) = status_str.parse::<RunStatus>() {
+                statuses.insert(id, status);
+            }
+        }
+
+        Ok(statuses)
+    }
+
     /// Link a routine run to a dispatched job.
     pub async fn link_routine_run_to_job(
         &self,

src/llm/anthropic_oauth.rs

@@ -575,6 +575,7 @@ fn extract_response_content(response: &AnthropicResponse) -> (Option<String>, Ve
                     id: id.clone(),
                     name: name.clone(),
                     arguments: input.clone(),
+                    reasoning: None,
                 });
             }
         }
@@ -623,6 +624,7 @@ mod tests {
             id: "call_1".to_string(),
             name: "search".to_string(),
             arguments: serde_json::json!({"q": "test"}),
+            reasoning: None,
         }];
         let messages = vec![
             ChatMessage::user("Search for test"),

src/llm/bedrock.rs

@@ -522,6 +522,7 @@ fn extract_content_blocks(
                     id: tu.tool_use_id().to_string(),
                     name: tu.name().to_string(),
                     arguments: document_to_json(tu.input()),
+                    reasoning: None,
                 });
             }
             // Ignore reasoning, citations, images, etc.
@@ -759,11 +760,13 @@ mod tests {
             id: "call_1".to_string(),
             name: "echo".to_string(),
             arguments: serde_json::json!({"text": "hi"}),
+            reasoning: None,
         };
         let tc2 = crate::llm::provider::ToolCall {
             id: "call_2".to_string(),
             name: "time".to_string(),
             arguments: serde_json::json!({}),
+            reasoning: None,
         };
 
         let messages = vec![
@@ -802,6 +805,7 @@ mod tests {
             id: "call_1".to_string(),
             name: "search".to_string(),
             arguments: serde_json::json!({"query": "test"}),
+            reasoning: None,
         };
 
         let messages = vec![
@@ -825,6 +829,7 @@ mod tests {
             id: "call_1".to_string(),
             name: "echo".to_string(),
             arguments: serde_json::json!({}),
+            reasoning: None,
         };
 
         let messages = vec![
@@ -989,11 +994,13 @@ mod tests {
             id: "call_abc".to_string(),
             name: "get_weather".to_string(),
             arguments: serde_json::json!({"city": "NYC"}),
+            reasoning: None,
         };
         let tc2 = crate::llm::provider::ToolCall {
             id: "call_def".to_string(),
             name: "get_time".to_string(),
             arguments: serde_json::json!({"tz": "EST"}),
+            reasoning: None,
         };
 
         let messages = vec![

src/llm/codex_chatgpt.rs

@@ -732,6 +732,7 @@ impl LlmProvider for CodexChatGptProvider {
                     id: tc.call_id,
                     name: tc.name,
                     arguments: args,
+                    reasoning: None,
                 }
             })
             .collect();
@@ -825,6 +826,7 @@ mod tests {
             id: "call_1".to_string(),
             name: "search".to_string(),
             arguments: json!({"query": "rust"}),
+            reasoning: None,
         };
         let msg = ChatMessage::assistant_with_tool_calls(Some("thinking...".into()), vec![tc]);
         let items = CodexChatGptProvider::message_to_input_items(&msg);

src/llm/gemini_oauth.rs

@@ -1898,6 +1898,7 @@ impl GeminiOauthProvider {
                         id,
                         name,
                         arguments: args,
+                        reasoning: None,
                     });
                 }
             }

src/llm/github_copilot.rs

@@ -107,14 +107,21 @@ impl GithubCopilotProvider {
         body: &impl Serialize,
     ) -> Result<R, LlmError> {
         let url = self.api_url();
-        // Map token exchange failures to RequestFailed (retryable) rather than
+        // Distinguish permanent auth errors (non-retryable) from transient
-        // AuthFailed (non-retryable), since transient network errors during
+        // network failures (retryable) so RetryProvider handles them correctly.
-        // exchange should be retried by RetryProvider.
         let token = self.token_manager.get_token().await.map_err(|e| {
             tracing::warn!(error = %e, "Copilot: token exchange failed");
-            LlmError::RequestFailed {
+            match &e {
-                provider: "github_copilot".to_string(),
+                crate::llm::github_copilot_auth::GithubCopilotAuthError::AccessDenied
-                reason: format!("Token exchange failed: {e}"),
+                | crate::llm::github_copilot_auth::GithubCopilotAuthError::Expired => {
+                    LlmError::AuthFailed {
+                        provider: "github_copilot".to_string(),
+                    }
+                }
+                _ => LlmError::RequestFailed {
+                    provider: "github_copilot".to_string(),
+                    reason: format!("Token exchange failed: {e}"),
+                },
             }
         })?;
 
@@ -157,54 +164,14 @@ impl GithubCopilotProvider {
             );
 
             if status.as_u16() == 401 {
-                // Invalidate the cached session token and retry once with a
+                // Invalidate the cached session token so the next attempt
-                // fresh exchange — stale tokens are the most common 401 cause.
+                // (driven by RetryProvider) gets a fresh one. We don't retry
-                tracing::warn!("Copilot: 401 Unauthorized — invalidating session token, retrying");
+                // inline to avoid nested retries with the outer RetryProvider.
+                tracing::warn!("Copilot: 401 Unauthorized — invalidating session token for retry");
                 self.token_manager.invalidate().await;
-                let fresh = self.token_manager.get_token().await.map_err(|e| {
+                return Err(LlmError::RequestFailed {
-                    tracing::warn!(error = %e, "Copilot: re-exchange after 401 failed");
-                    LlmError::RequestFailed {
-                        provider: "github_copilot".to_string(),
-                        reason: format!("Token re-exchange after 401 failed: {e}"),
-                    }
-                })?;
-                let mut retry_req = self
-                    .client
-                    .post(&url)
-                    .bearer_auth(fresh.expose_secret())
-                    .header("Content-Type", "application/json");
-                for (key, value) in &self.extra_headers {
-                    retry_req = retry_req.header(key.as_str(), value.as_str());
-                }
-                let retry =
-                    retry_req
-                        .json(body)
-                        .send()
-                        .await
-                        .map_err(|e| LlmError::RequestFailed {
-                            provider: "github_copilot".to_string(),
-                            reason: format!("Retry after 401 failed: {e}"),
-                        })?;
-                if retry.status().is_success() {
-                    let text = retry.text().await.map_err(|e| LlmError::RequestFailed {
-                        provider: "github_copilot".to_string(),
-                        reason: format!("Failed to read retry response body: {e}"),
-                    })?;
-                    return serde_json::from_str(&text).map_err(|e| {
-                        let truncated = crate::agent::truncate_for_preview(&text, 512);
-                        LlmError::InvalidResponse {
-                            provider: "github_copilot".to_string(),
-                            reason: format!("JSON parse error: {e}. Raw: {truncated}"),
-                        }
-                    });
-                }
-                let retry_status = retry.status();
-                tracing::warn!(
-                    status = %retry_status,
-                    "Copilot: 401 retry also failed"
-                );
-                return Err(LlmError::AuthFailed {
                     provider: "github_copilot".to_string(),
+                    reason: "HTTP 401 Unauthorized".to_string(),
                 });
             }
             if status.as_u16() == 429 {
@@ -629,6 +596,7 @@ fn extract_choice_content(choice: &OpenAiChoice) -> (Option<String>, Vec<ToolCal
                     name: tc.function.name.clone(),
                     arguments: serde_json::from_str(&tc.function.arguments)
                         .unwrap_or(serde_json::Value::Object(serde_json::Map::new())),
+                    reasoning: None,
                 })
                 .collect()
         })
@@ -661,6 +629,7 @@ mod tests {
             id: "call_1".to_string(),
             name: "search".to_string(),
             arguments: serde_json::json!({"q": "test"}),
+            reasoning: None,
         }];
         let messages = vec![
             ChatMessage::user("Search"),

src/llm/nearai_chat.rs

@@ -463,8 +463,15 @@ impl LlmProvider for NearAiChatProvider {
         let model = req.model.unwrap_or_else(|| self.active_model_name());
         let mut raw_messages = req.messages;
         crate::llm::provider::sanitize_tool_messages(&mut raw_messages);
-        let messages: Vec<ChatCompletionMessage> =
+        let raw: Vec<ChatCompletionMessage> = raw_messages.into_iter().map(|m| m.into()).collect();
-            raw_messages.into_iter().map(|m| m.into()).collect();
+
+        // NEAR AI rejects `role:"tool"` messages even on text-only completion paths.
+        // Apply the same flattening used by complete_with_tools().
+        let messages = if self.flatten_tool_messages {
+            flatten_tool_messages(raw)
+        } else {
+            raw
+        };
 
         let request = ChatCompletionRequest {
             model,
@@ -580,6 +587,7 @@ impl LlmProvider for NearAiChatProvider {
                     id: tc.id,
                     name: tc.function.name,
                     arguments,
+                    reasoning: None,
                 }
             })
             .collect();
@@ -1173,11 +1181,13 @@ mod tests {
                 id: "call_1".to_string(),
                 name: "list_issues".to_string(),
                 arguments: serde_json::json!({"owner": "foo", "repo": "bar"}),
+                reasoning: None,
             },
             ToolCall {
                 id: "call_2".to_string(),
                 name: "search".to_string(),
                 arguments: serde_json::json!({"query": "test"}),
+                reasoning: None,
             },
         ];
 
@@ -1210,6 +1220,7 @@ mod tests {
             id: "call_1".to_string(),
             name: "test".to_string(),
             arguments: serde_json::json!({"key": "value"}),
+            reasoning: None,
         };
         let msg = ChatMessage::assistant_with_tool_calls(None, vec![tc]);
         let chat_msg: ChatCompletionMessage = msg.into();
@@ -1453,6 +1464,7 @@ mod tests {
                     id: tc.id,
                     name: tc.function.name,
                     arguments,
+                    reasoning: None,
                 }
             })
             .collect();
@@ -1502,6 +1514,7 @@ mod tests {
                     id: tc.id,
                     name: tc.function.name,
                     arguments,
+                    reasoning: None,
                 }
             })
             .collect();
@@ -2124,6 +2137,7 @@ mod tests {
                 id: "call_1".to_string(),
                 name: "test".to_string(),
                 arguments: serde_json::json!({}),
+                reasoning: None,
             }],
         );
         let chat_msg: ChatCompletionMessage = msg.into();
@@ -2193,6 +2207,65 @@ mod tests {
         assert_eq!(deserialized.function.arguments, r#"{"city":"London"}"#);
     }
 
+    // -- flatten_tool_messages in complete() path ----------------------------
+
+    #[test]
+    fn test_flatten_applied_on_text_only_path() {
+        // Verify that flatten_tool_messages converts tool-role messages to user
+        // messages (mirrors the complete_with_tools path).
+        let messages = vec![
+            ChatCompletionMessage {
+                role: "user".to_string(),
+                content: Some(MessageContent::Text("run it".to_string())),
+                tool_call_id: None,
+                name: None,
+                tool_calls: None,
+            },
+            ChatCompletionMessage {
+                role: "tool".to_string(),
+                content: Some(MessageContent::Text("ok".to_string())),
+                tool_call_id: Some("call_1".to_string()),
+                name: Some("run_cmd".to_string()),
+                tool_calls: None,
+            },
+        ];
+        let flattened = flatten_tool_messages(messages);
+        assert_eq!(flattened.len(), 2);
+        assert_eq!(flattened[1].role, "user");
+        let text = flattened[1]
+            .content
+            .as_ref()
+            .and_then(|c| c.as_text())
+            .unwrap();
+        assert!(text.contains("run_cmd"), "should reference tool name");
+        assert!(text.contains("ok"), "should include tool result");
+    }
+
+    #[test]
+    fn test_no_flatten_when_no_tool_messages() {
+        // When there are no tool-role messages, flatten_tool_messages is a no-op.
+        let messages = vec![
+            ChatCompletionMessage {
+                role: "user".to_string(),
+                content: Some(MessageContent::Text("hi".to_string())),
+                tool_call_id: None,
+                name: None,
+                tool_calls: None,
+            },
+            ChatCompletionMessage {
+                role: "assistant".to_string(),
+                content: Some(MessageContent::Text("hello".to_string())),
+                tool_call_id: None,
+                name: None,
+                tool_calls: None,
+            },
+        ];
+        let result = flatten_tool_messages(messages);
+        // No tool messages → unchanged roles
+        assert_eq!(result[0].role, "user");
+        assert_eq!(result[1].role, "assistant");
+    }
+
     // -- api_url edge cases ---------------------------------------------------
 
     #[test]

src/llm/openai_codex_provider.rs

@@ -625,6 +625,7 @@ fn parse_sse_response(body: &str) -> Result<ParsedResponse, LlmError> {
                                 id: state.call_id,
                                 name: state.name,
                                 arguments,
+                                reasoning: None,
                             });
                         } else {
                             // Fallback: extract directly from the item
@@ -650,6 +651,7 @@ fn parse_sse_response(body: &str) -> Result<ParsedResponse, LlmError> {
                                 id: call_id,
                                 name,
                                 arguments,
+                                reasoning: None,
                             });
                         }
                     }
@@ -727,6 +729,7 @@ fn parse_sse_response(body: &str) -> Result<ParsedResponse, LlmError> {
                 id: state.call_id,
                 name: state.name,
                 arguments,
+                reasoning: None,
             });
         }
     }
@@ -822,11 +825,13 @@ mod tests {
                 id: "call_1".to_string(),
                 name: "search".to_string(),
                 arguments: serde_json::json!({"query": "test"}),
+                reasoning: None,
             },
             ToolCall {
                 id: "call_2".to_string(),
                 name: "read".to_string(),
                 arguments: serde_json::json!({"path": "/tmp"}),
+                reasoning: None,
             },
         ];
         let msg =

src/llm/provider.rs

@@ -231,6 +231,10 @@ pub struct ToolCall {
     pub id: String,
     pub name: String,
     pub arguments: serde_json::Value,
+    /// Optional reasoning for why this tool was chosen — supplied by the provider
+    /// or derived from the shared response content as a fallback.
+    #[serde(default, skip_serializing_if = "Option::is_none")]
+    pub reasoning: Option<String>,
 }
 
 /// Generate a tool-call ID that satisfies all providers.
@@ -637,6 +641,7 @@ mod tests {
             id: "call_1".to_string(),
             name: "echo".to_string(),
             arguments: serde_json::json!({}),
+            reasoning: None,
         };
         let mut messages = vec![
             ChatMessage::user("hello"),
@@ -680,6 +685,7 @@ mod tests {
             id: "call_1".to_string(),
             name: "echo".to_string(),
             arguments: serde_json::json!({}),
+            reasoning: None,
         };
         let mut messages = vec![
             ChatMessage::user("test"),
@@ -705,11 +711,13 @@ mod tests {
             id: "call_sel_1".to_string(),
             name: "search".to_string(),
             arguments: serde_json::json!({"q": "test"}),
+            reasoning: None,
         };
         let tc2 = ToolCall {
             id: "call_sel_2".to_string(),
             name: "http".to_string(),
             arguments: serde_json::json!({"url": "https://example.com"}),
+            reasoning: None,
         };
         let mut messages = vec![
             ChatMessage::system("You are a helpful assistant."),

src/llm/reasoning.rs

@@ -525,17 +525,35 @@ impl Reasoning {
 
         let response = self.llm.complete_with_tools(request).await?;
 
-        let reasoning = response.content.unwrap_or_default();
+        let shared_reasoning = response
+            .content
+            .map(|c| {
+                let pre_truncated = truncate_at_tool_tags(&c);
+                clean_response(&pre_truncated)
+            })
+            .unwrap_or_default();
 
         let selections: Vec<ToolSelection> = response
             .tool_calls
             .into_iter()
-            .map(|tool_call| ToolSelection {
+            .map(|tool_call| {
-                tool_name: tool_call.name,
+                // Prefer per-tool reasoning if the provider supplied it,
-                parameters: tool_call.arguments,
+                // otherwise fall back to the shared response content.
-                reasoning: reasoning.clone(),
+                let rationale = tool_call
-                alternatives: vec![],
+                    .reasoning
-                tool_call_id: tool_call.id,
+                    .map(|r| {
+                        let pre_truncated = truncate_at_tool_tags(&r);
+                        clean_response(&pre_truncated)
+                    })
+                    .filter(|r| !r.trim().is_empty())
+                    .unwrap_or_else(|| shared_reasoning.clone());
+                ToolSelection {
+                    tool_name: tool_call.name,
+                    parameters: tool_call.arguments,
+                    reasoning: rationale,
+                    alternatives: vec![],
+                    tool_call_id: tool_call.id,
+                }
             })
             .collect();
 
@@ -664,13 +682,36 @@ Respond in JSON format:
 
             // If there were tool calls, return them for execution
             if !response.tool_calls.is_empty() {
+                let narrative = response.content.map(|c| {
+                    let pre_truncated = truncate_at_tool_tags(&c);
+                    clean_response(&pre_truncated)
+                });
+                // Populate per-tool reasoning from the shared narrative when the
+                // provider did not supply per-tool rationale.
+                let tool_calls: Vec<ToolCall> = response
+                    .tool_calls
+                    .into_iter()
+                    .map(|mut tc| {
+                        if tc.reasoning.as_ref().is_none_or(|r| r.trim().is_empty()) {
+                            tc.reasoning = narrative.as_ref().filter(|n| !n.is_empty()).cloned();
+                        } else {
+                            // Clean provider-supplied per-tool reasoning the same way
+                            // we clean the shared narrative (strip thinking/tool tags).
+                            tc.reasoning = tc
+                                .reasoning
+                                .map(|r| {
+                                    let pre_truncated = truncate_at_tool_tags(&r);
+                                    clean_response(&pre_truncated)
+                                })
+                                .filter(|r| !r.trim().is_empty());
+                        }
+                        tc
+                    })
+                    .collect();
                 return Ok(RespondOutput {
                     result: RespondResult::ToolCalls {
-                        tool_calls: response.tool_calls,
+                        tool_calls,
-                        content: response.content.map(|c| {
+                        content: narrative,
-                            let pre_truncated = truncate_at_tool_tags(&c);
-                            clean_response(&pre_truncated)
-                        }),
                     },
                     usage,
                 });
@@ -1350,6 +1391,7 @@ fn recover_tool_calls_from_content(
                     ),
                     name: name.to_string(),
                     arguments,
+                    reasoning: None,
                 });
                 continue;
             }
@@ -1364,6 +1406,7 @@ fn recover_tool_calls_from_content(
                     ),
                     name: name.to_string(),
                     arguments: serde_json::Value::Object(Default::default()),
+                    reasoning: None,
                 });
             }
         }
@@ -1401,6 +1444,7 @@ fn recover_tool_calls_from_content(
                         ),
                         name: name.to_string(),
                         arguments,
+                        reasoning: None,
                     });
                     remaining = &args_start[bracket_end + 1..];
                     continue;
@@ -1412,6 +1456,7 @@ fn recover_tool_calls_from_content(
                 id: super::provider::generate_tool_call_id(calls.len(), RECOVERED_TOOL_CALL_SEED),
                 name: name.to_string(),
                 arguments: serde_json::Value::Object(Default::default()),
+                reasoning: None,
             });
             remaining = after_name;
         }
@@ -3145,4 +3190,32 @@ That's my plan."#;
             "Text <function_call>{}</function_call> middle "
         );
     }
+
+    /// Verify that reasoning normalization strips thinking tags and tool tags
+    /// from per-tool reasoning, matching the cleaning applied to shared reasoning.
+    #[test]
+    fn test_reasoning_normalization_strips_thinking_tags() {
+        let raw = "<thinking>Let me consider...</thinking>Search memory for prior context";
+        let pre_truncated = truncate_at_tool_tags(raw);
+        let cleaned = clean_response(&pre_truncated);
+        assert!(!cleaned.contains("<thinking>"));
+        assert!(cleaned.contains("Search memory"));
+    }
+
+    #[test]
+    fn test_reasoning_normalization_strips_tool_tags() {
+        let raw = "Calling search <tool_call>{\"name\": \"search\"}";
+        let pre_truncated = truncate_at_tool_tags(raw);
+        let cleaned = clean_response(&pre_truncated);
+        assert!(!cleaned.contains("<tool_call>"));
+        assert!(cleaned.contains("Calling search"));
+    }
+
+    #[test]
+    fn test_reasoning_normalization_empty_after_cleaning() {
+        let raw = "<thinking>internal only</thinking>";
+        let pre_truncated = truncate_at_tool_tags(raw);
+        let cleaned = clean_response(&pre_truncated);
+        assert!(cleaned.trim().is_empty());
+    }
 }

src/llm/rig_adapter.rs

@@ -490,6 +490,7 @@ fn extract_response(
                     id: tc.id.clone(),
                     name: tc.function.name.clone(),
                     arguments: tc.function.arguments.clone(),
+                    reasoning: None,
                 });
             }
             // Reasoning and Image variants are not mapped to IronClaw types
@@ -880,6 +881,7 @@ mod tests {
             id: "Xt7mK9pQ2".to_string(),
             name: "search".to_string(),
             arguments: serde_json::json!({"query": "test"}),
+            reasoning: None,
         };
         let msg = ChatMessage::assistant_with_tool_calls(Some("thinking".to_string()), vec![tc]);
         let messages = vec![msg];
@@ -997,6 +999,7 @@ mod tests {
             id: "".to_string(),
             name: "search".to_string(),
             arguments: serde_json::json!({"query": "test"}),
+            reasoning: None,
         };
         let messages = vec![ChatMessage::assistant_with_tool_calls(None, vec![tc])];
         let (_preamble, history) = convert_messages(&messages);
@@ -1028,6 +1031,7 @@ mod tests {
             id: "   ".to_string(),
             name: "search".to_string(),
             arguments: serde_json::json!({"query": "test"}),
+            reasoning: None,
         };
         let messages = vec![ChatMessage::assistant_with_tool_calls(None, vec![tc])];
         let (_preamble, history) = convert_messages(&messages);
@@ -1061,6 +1065,7 @@ mod tests {
             id: "".to_string(),
             name: "search".to_string(),
             arguments: serde_json::json!({"query": "test"}),
+            reasoning: None,
         };
         let assistant_msg = ChatMessage::assistant_with_tool_calls(None, vec![tc]);
         let tool_result_msg = ChatMessage {
@@ -1380,11 +1385,13 @@ mod tests {
             id: "call_a".to_string(),
             name: "search".to_string(),
             arguments: serde_json::json!({"q": "rust"}),
+            reasoning: None,
         };
         let tc2 = IronToolCall {
             id: "call_b".to_string(),
             name: "fetch".to_string(),
             arguments: serde_json::json!({"url": "https://example.com"}),
+            reasoning: None,
         };
         let assistant = ChatMessage::assistant_with_tool_calls(None, vec![tc1, tc2]);
         let result_a = ChatMessage::tool_result("call_a", "search", "search results");

src/main.rs

@@ -142,6 +142,11 @@ async fn async_main() -> anyhow::Result<()> {
             init_cli_tracing();
             return ironclaw::cli::run_logs_command(logs_cmd.clone(), cli.config.as_deref()).await;
         }
+        Some(Command::Models(models_cmd)) => {
+            init_cli_tracing();
+            return ironclaw::cli::run_models_command(models_cmd.clone(), cli.config.as_deref())
+                .await;
+        }
         Some(Command::Doctor) => {
             init_cli_tracing();
             return ironclaw::cli::run_doctor_command().await;
@@ -584,15 +589,47 @@ async fn async_main() -> anyhow::Result<()> {
     // ── Gateway channel ────────────────────────────────────────────────
 
     let mut gateway_url: Option<String> = None;
-    let mut sse_sender: Option<
+    let mut sse_manager: Option<std::sync::Arc<ironclaw::channels::web::sse::SseManager>> = None;
-        tokio::sync::broadcast::Sender<ironclaw::channels::web::types::SseEvent>,
-    > = None;
     if let Some(ref gw_config) = config.channels.gateway {
-        let mut gw =
+        // Build multi-user auth state if user_tokens is configured, else single-user.
-            GatewayChannel::new(gw_config.clone()).with_llm_provider(Arc::clone(&components.llm));
+        let mut gw = if let Some(ref user_tokens) = gw_config.user_tokens {
+            use ironclaw::channels::web::auth::{MultiAuthState, UserIdentity};
+            let tokens = user_tokens
+                .iter()
+                .map(|(token, cfg)| {
+                    (
+                        token.clone(),
+                        UserIdentity {
+                            user_id: cfg.user_id.clone(),
+                            workspace_read_scopes: cfg.workspace_read_scopes.clone(),
+                        },
+                    )
+                })
+                .collect();
+            let auth = MultiAuthState::multi(tokens);
+            GatewayChannel::new_multi_auth(gw_config.clone(), auth)
+        } else {
+            GatewayChannel::new(gw_config.clone())
+        };
+        gw = gw.with_owner_scope(config.owner_id.clone());
+        gw = gw.with_llm_provider(Arc::clone(&components.llm));
         if let Some(ref ws) = components.workspace {
             gw = gw.with_workspace(Arc::clone(ws));
         }
+        // Create per-user workspace pool for multi-user mode.
+        if let Some(ref db) = components.db {
+            let emb_cache_config = ironclaw::workspace::EmbeddingCacheConfig {
+                max_entries: config.embeddings.cache_size,
+            };
+            let pool = Arc::new(ironclaw::channels::web::server::WorkspacePool::new(
+                Arc::clone(db),
+                components.embeddings.clone(),
+                emb_cache_config,
+                config.search.clone(),
+                config.workspace.clone(),
+            ));
+            gw = gw.with_workspace_pool(pool);
+        }
         gw = gw.with_session_manager(Arc::clone(&session_manager));
         gw = gw.with_log_broadcaster(Arc::clone(&log_broadcaster));
         gw = gw.with_log_level_handle(Arc::clone(&log_level_handle));
@@ -643,8 +680,12 @@ async fn async_main() -> anyhow::Result<()> {
                 let mut rx = tx.subscribe();
                 let gw_state = Arc::clone(gw.state());
                 tokio::spawn(async move {
-                    while let Ok((_job_id, event)) = rx.recv().await {
+                    while let Ok((_job_id, user_id, event)) = rx.recv().await {
-                        gw_state.sse.broadcast(event);
+                        if user_id.is_empty() {
+                            gw_state.sse.broadcast(event);
+                        } else {
+                            gw_state.sse.broadcast_for_user(&user_id, event);
+                        }
                     }
                 });
             }
@@ -686,7 +727,7 @@ async fn async_main() -> anyhow::Result<()> {
         // Capture SSE sender and routine engine slot before moving gw into channels.
         // IMPORTANT: This must come after all `with_*` calls since `rebuild_state`
         // creates a new SseManager, which would orphan this sender.
-        sse_sender = Some(gw.state().sse.sender());
+        sse_manager = Some(Arc::clone(&gw.state().sse));
         channel_names.push("gateway".to_string());
         channels.add(Box::new(gw)).await;
     }
@@ -749,6 +790,14 @@ async fn async_main() -> anyhow::Result<()> {
         .register_message_tools(Arc::clone(&channels), components.extension_manager.clone())
         .await;
 
+    // Default user ID for extension operations (single-user mode).
+    let ext_user_id = config
+        .channels
+        .gateway
+        .as_ref()
+        .map(|g| g.user_id.clone())
+        .unwrap_or_else(|| "default".to_string());
+
     // Wire up channel runtime for hot-activation of WASM channels.
     if let Some(ref ext_mgr) = components.extension_manager
         && let Some((rt, ps, router)) = wasm_channel_runtime_state.take()
@@ -769,12 +818,14 @@ async fn async_main() -> anyhow::Result<()> {
 
         // Auto-activate WASM channels that were active in a previous session.
         // Relay channels are handled separately below via restore_relay_channels().
-        let persisted = ext_mgr.load_persisted_active_channels().await;
+        let persisted = ext_mgr.load_persisted_active_channels(&ext_user_id).await;
         for name in &persisted {
-            if active_at_startup.contains(name) || ext_mgr.is_relay_channel(name).await {
+            if active_at_startup.contains(name)
+                || ext_mgr.is_relay_channel(name, &ext_user_id).await
+            {
                 continue;
             }
-            match ext_mgr.activate(name).await {
+            match ext_mgr.activate(name, &ext_user_id).await {
                 Ok(result) => {
                     tracing::debug!(
                         channel = %name,
@@ -799,14 +850,14 @@ async fn async_main() -> anyhow::Result<()> {
         ext_mgr
             .set_relay_channel_manager(Arc::clone(&channels))
             .await;
-        ext_mgr.restore_relay_channels().await;
+        ext_mgr.restore_relay_channels(&ext_user_id).await;
     }
 
     // Wire SSE sender into extension manager for broadcasting status events.
     if let Some(ref ext_mgr) = components.extension_manager
-        && let Some(ref sender) = sse_sender
+        && let Some(ref sse) = sse_manager
     {
-        ext_mgr.set_sse_sender(sender.clone()).await;
+        ext_mgr.set_sse_sender(Arc::clone(sse)).await;
     }
 
     // Snapshot memory for trace recording before the agent starts
@@ -844,7 +895,7 @@ async fn async_main() -> anyhow::Result<()> {
         skills_config: config.skills.clone(),
         hooks: components.hooks,
         cost_guard: components.cost_guard,
-        sse_tx: sse_sender,
+        sse_tx: sse_manager,
         http_interceptor,
         transcription: config.transcription.create_provider().map(|p| {
             Arc::new(ironclaw::llm::transcription::TranscriptionMiddleware::new(
@@ -862,6 +913,7 @@ async fn async_main() -> anyhow::Result<()> {
             ironclaw::agent::routine_engine::SandboxReadiness::DockerUnavailable
         },
         builder: components.builder,
+        llm_backend: config.llm.backend.clone(),
     };
 
     let channels_for_warnings = Arc::clone(&channels);

src/orchestrator/api.rs

@@ -14,7 +14,7 @@ use serde::{Deserialize, Serialize};
 use tokio::sync::{Mutex, broadcast};
 use uuid::Uuid;
 
-use crate::channels::web::types::SseEvent;
+use crate::channels::web::types::ToolDecisionDto;
 use crate::db::Database;
 use crate::llm::{CompletionRequest, LlmProvider, ToolCompletionRequest};
 use crate::orchestrator::auth::{TokenStore, worker_auth_middleware};
@@ -25,6 +25,7 @@ use crate::worker::api::{
     CompletionReport, CredentialResponse, JobDescription, ProxyCompletionRequest,
     ProxyCompletionResponse, ProxyToolCompletionRequest, ProxyToolCompletionResponse, StatusUpdate,
 };
+use ironclaw_common::AppEvent;
 
 /// A follow-up prompt queued for a Claude Code bridge.
 #[derive(Debug, Clone, Serialize, Deserialize)]
@@ -40,7 +41,8 @@ pub struct OrchestratorState {
     pub job_manager: Arc<ContainerJobManager>,
     pub token_store: TokenStore,
     /// Broadcast channel for job events (consumed by the web gateway SSE).
-    pub job_event_tx: Option<broadcast::Sender<(Uuid, SseEvent)>>,
+    /// Tuple: (job_id, user_id, event).
+    pub job_event_tx: Option<broadcast::Sender<(Uuid, String, AppEvent)>>,
     /// Buffered follow-up prompts for sandbox jobs, keyed by job_id.
     pub prompt_queue: Arc<Mutex<HashMap<Uuid, VecDeque<PendingPrompt>>>>,
     /// Database handle for persisting job events.
@@ -49,6 +51,9 @@ pub struct OrchestratorState {
     pub secrets_store: Option<Arc<dyn SecretsStore + Send + Sync>>,
     /// User ID for secret lookups (single-tenant, typically "default").
     pub user_id: String,
+    /// In-memory cache of job_id → user_id for SSE scoping. Populated when
+    /// sandbox jobs are created, avoiding a DB round-trip on every job event.
+    pub job_owner_cache: Arc<std::sync::RwLock<HashMap<Uuid, String>>>,
 }
 
 /// The orchestrator's internal API server.
@@ -273,10 +278,10 @@ async fn job_event_handler(
         });
     }
 
-    // Convert to SSE event and broadcast
+    // Convert to app event and broadcast
     let job_id_str = job_id.to_string();
-    let sse_event = match payload.event_type.as_str() {
+    let app_event = match payload.event_type.as_str() {
-        "message" => SseEvent::JobMessage {
+        "message" => AppEvent::JobMessage {
             job_id: job_id_str,
             role: payload
                 .data
@@ -291,7 +296,7 @@ async fn job_event_handler(
                 .unwrap_or("")
                 .to_string(),
         },
-        "tool_use" => SseEvent::JobToolUse {
+        "tool_use" => AppEvent::JobToolUse {
             job_id: job_id_str,
             tool_name: payload
                 .data
@@ -305,7 +310,7 @@ async fn job_event_handler(
                 .cloned()
                 .unwrap_or(serde_json::Value::Null),
         },
-        "tool_result" => SseEvent::JobToolResult {
+        "tool_result" => AppEvent::JobToolResult {
             job_id: job_id_str,
             tool_name: payload
                 .data
@@ -320,7 +325,7 @@ async fn job_event_handler(
                 .unwrap_or("")
                 .to_string(),
         },
-        "result" => SseEvent::JobResult {
+        "result" => AppEvent::JobResult {
             job_id: job_id_str,
             status: payload
                 .data
@@ -340,7 +345,21 @@ async fn job_event_handler(
             // gain context/memory tracking capabilities.
             fallback_deliverable: payload.data.get("fallback_deliverable").cloned(),
         },
-        _ => SseEvent::JobStatus {
+        "reasoning" => {
+            let narrative = payload
+                .data
+                .get("narrative")
+                .and_then(|v| v.as_str())
+                .unwrap_or("")
+                .to_string();
+            let decisions = ToolDecisionDto::from_json_array(&payload.data["decisions"]);
+            AppEvent::JobReasoning {
+                job_id: job_id_str,
+                narrative,
+                decisions,
+            }
+        }
+        _ => AppEvent::JobStatus {
             job_id: job_id_str,
             message: payload
                 .data
@@ -351,9 +370,45 @@ async fn job_event_handler(
         },
     };
 
-    // Broadcast via the channel (if configured)
+    // Broadcast via the channel (if configured).
+    // Look up the job owner from the in-memory cache (populated at job creation).
     if let Some(ref tx) = state.job_event_tx {
-        let _ = tx.send((job_id, sse_event));
+        let cached_uid = state
+            .job_owner_cache
+            .read()
+            .unwrap_or_else(|e| e.into_inner())
+            .get(&job_id)
+            .cloned();
+
+        let user_id = match cached_uid {
+            Some(uid) => uid,
+            None => {
+                // Cache miss: fall back to DB lookup and populate cache.
+                let uid = match state.store.as_ref() {
+                    Some(store) => store
+                        .get_sandbox_job(job_id)
+                        .await
+                        .ok()
+                        .flatten()
+                        .map(|j| j.user_id),
+                    None => None,
+                };
+                if let Some(ref uid) = uid {
+                    state
+                        .job_owner_cache
+                        .write()
+                        .unwrap_or_else(|e| e.into_inner())
+                        .insert(job_id, uid.clone());
+                }
+                uid.unwrap_or_default()
+            }
+        };
+
+        if user_id.is_empty() {
+            let _ = tx.send((job_id, String::new(), app_event));
+        } else {
+            let _ = tx.send((job_id, user_id, app_event));
+        }
     }
 
     Ok(StatusCode::OK)
@@ -480,6 +535,7 @@ mod tests {
             store: None,
             secrets_store: None,
             user_id: "default".to_string(),
+            job_owner_cache: Arc::new(std::sync::RwLock::new(HashMap::new())),
         }
     }
 
@@ -709,6 +765,7 @@ mod tests {
             store: None,
             secrets_store: Some(secrets_store),
             user_id: "default".to_string(),
+            job_owner_cache: Arc::new(std::sync::RwLock::new(HashMap::new())),
         };
 
         let router = OrchestratorApi::router(state);
@@ -744,6 +801,7 @@ mod tests {
             store: None,
             secrets_store: None,
             user_id: "default".to_string(),
+            job_owner_cache: Arc::new(std::sync::RwLock::new(HashMap::new())),
         };
 
         let job_id = Uuid::new_v4();
@@ -769,10 +827,12 @@ mod tests {
         let resp = router.oneshot(req).await.unwrap();
         assert_eq!(resp.status(), StatusCode::OK);
 
-        let (recv_id, event) = rx.recv().await.unwrap();
+        let (recv_id, recv_uid, event) = rx.recv().await.unwrap();
         assert_eq!(recv_id, job_id);
+        // No store configured, so user_id falls back to empty string.
+        assert_eq!(recv_uid, "");
         match event {
-            SseEvent::JobMessage {
+            AppEvent::JobMessage {
                 job_id: jid,
                 role,
                 content,
@@ -799,6 +859,7 @@ mod tests {
             store: None,
             secrets_store: None,
             user_id: "default".to_string(),
+            job_owner_cache: Arc::new(std::sync::RwLock::new(HashMap::new())),
         };
 
         let job_id = Uuid::new_v4();
@@ -824,9 +885,9 @@ mod tests {
         let resp = router.oneshot(req).await.unwrap();
         assert_eq!(resp.status(), StatusCode::OK);
 
-        let (_recv_id, event) = rx.recv().await.unwrap();
+        let (_recv_id, _recv_uid, event) = rx.recv().await.unwrap();
         match event {
-            SseEvent::JobToolUse { tool_name, .. } => {
+            AppEvent::JobToolUse { tool_name, .. } => {
                 assert_eq!(tool_name, "shell");
             }
             other => panic!("Expected JobToolUse, got {:?}", other),
@@ -847,6 +908,7 @@ mod tests {
             store: None,
             secrets_store: None,
             user_id: "default".to_string(),
+            job_owner_cache: Arc::new(std::sync::RwLock::new(HashMap::new())),
         };
 
         let job_id = Uuid::new_v4();
@@ -869,9 +931,9 @@ mod tests {
         let resp = router.oneshot(req).await.unwrap();
         assert_eq!(resp.status(), StatusCode::OK);
 
-        let (_recv_id, event) = rx.recv().await.unwrap();
+        let (_recv_id, _recv_uid, event) = rx.recv().await.unwrap();
         // Unknown event types fall through to JobStatus
-        assert!(matches!(event, SseEvent::JobStatus { .. }));
+        assert!(matches!(event, AppEvent::JobStatus { .. }));
     }
 
     // -- Status update test --

src/orchestrator/mod.rs

@@ -46,10 +46,10 @@ use std::sync::Arc;
 use tokio::sync::{Mutex, broadcast};
 use uuid::Uuid;
 
-use crate::channels::web::types::SseEvent;
 use crate::db::Database;
 use crate::llm::LlmProvider;
 use crate::secrets::SecretsStore;
+use ironclaw_common::AppEvent;
 
 /// Resolve the orchestrator port from the `ORCHESTRATOR_PORT` environment
 /// variable, falling back to 50051.
@@ -63,7 +63,7 @@ fn resolve_orchestrator_port() -> u16 {
 /// Result of orchestrator setup, containing all handles needed by the agent.
 pub struct OrchestratorSetup {
     pub container_job_manager: Option<Arc<ContainerJobManager>>,
-    pub job_event_tx: Option<broadcast::Sender<(Uuid, SseEvent)>>,
+    pub job_event_tx: Option<broadcast::Sender<(Uuid, String, AppEvent)>>,
     pub prompt_queue: Arc<Mutex<HashMap<Uuid, VecDeque<api::PendingPrompt>>>>,
     pub docker_status: crate::sandbox::DockerStatus,
 }
@@ -134,6 +134,7 @@ pub async fn setup_orchestrator(
             store: db.cloned(),
             secrets_store: secrets_store.cloned(),
             user_id: "default".to_string(),
+            job_owner_cache: Arc::new(std::sync::RwLock::new(std::collections::HashMap::new())),
         };
 
         tokio::spawn(async move {

src/settings.rs

@@ -1297,6 +1297,92 @@ mod tests {
         assert_eq!(loaded.heartbeat.interval_secs, 900);
     }
 
+    /// Regression: /model writes a single key ("selected_model") to the DB via
+    /// set_setting(). On restart, get_all_settings() returns ALL keys including
+    /// wizard-written defaults. The single-key update must survive the full
+    /// from_db_map() round trip.
+    #[test]
+    fn db_single_key_model_update_survives_roundtrip() {
+        // Step 1: Wizard writes full settings to DB (including selected_model
+        // from initial setup).
+        let wizard_settings = Settings {
+            llm_backend: Some("nearai".to_string()),
+            selected_model: Some("old-wizard-model".to_string()),
+            ..Default::default()
+        };
+        let mut db: std::collections::HashMap<String, serde_json::Value> =
+            wizard_settings.to_db_map();
+
+        // Step 2: User runs /model new-model — persist_selected_model writes
+        // a single key, overwriting the wizard value.
+        db.insert(
+            "selected_model".to_string(),
+            serde_json::Value::String("new-model".to_string()),
+        );
+
+        // Step 3: On restart, from_db_map() rebuilds Settings from the full
+        // DB map.
+        let restored = Settings::from_db_map(&db);
+        assert_eq!(
+            restored.selected_model,
+            Some("new-model".to_string()),
+            "/model change must survive DB round trip"
+        );
+    }
+
+    /// Regression: TOML overlay must not clobber a DB-persisted selected_model
+    /// when the TOML file matches the DB. This is the normal case after /model
+    /// successfully writes to both DB and TOML.
+    #[test]
+    fn toml_overlay_preserves_matching_model() {
+        // DB settings with new model from /model command.
+        let mut db_settings = Settings {
+            llm_backend: Some("nearai".to_string()),
+            selected_model: Some("new-model".to_string()),
+            ..Default::default()
+        };
+
+        // TOML also updated by /model command to the same value.
+        let toml_settings = Settings {
+            selected_model: Some("new-model".to_string()),
+            ..Default::default()
+        };
+
+        db_settings.merge_from(&toml_settings);
+        assert_eq!(
+            db_settings.selected_model,
+            Some("new-model".to_string()),
+            "TOML overlay must not clobber matching model"
+        );
+    }
+
+    /// Regression: when /model updates DB but TOML write fails, a stale TOML
+    /// file would overwrite the DB value. This test documents the priority:
+    /// TOML > DB (by design). persist_selected_model MUST update the TOML.
+    #[test]
+    fn stale_toml_overwrites_db_model() {
+        // DB has the new model from /model.
+        let mut db_settings = Settings {
+            selected_model: Some("new-model".to_string()),
+            ..Default::default()
+        };
+
+        // TOML still has the old model (write failed or was not attempted).
+        let stale_toml = Settings {
+            selected_model: Some("old-model".to_string()),
+            ..Default::default()
+        };
+
+        db_settings.merge_from(&stale_toml);
+        // This documents the current priority: TOML wins over DB.
+        // The fix in persist_selected_model ensures TOML is always updated.
+        assert_eq!(
+            db_settings.selected_model,
+            Some("old-model".to_string()),
+            "TOML overlay has higher priority than DB (by design)"
+        );
+    }
+
     /// Regression test: /model command must persist selected_model to TOML config.
     /// Prior to the fix, `set_model()` only changed the in-memory provider and the
     /// choice was lost on restart.
@@ -1322,6 +1408,28 @@ mod tests {
         assert_eq!(reloaded.selected_model, Some("new-model".to_string()));
     }
 
+    /// Regression: /model must create config.toml when it doesn't exist, so the
+    /// model survives restarts. Previously the Ok(None) case was a no-op.
+    #[test]
+    fn toml_created_when_missing_for_model_persist() {
+        let dir = tempfile::tempdir().unwrap();
+        let path = dir.path().join("config.toml");
+
+        // No config.toml yet (fresh install, no wizard).
+        assert!(Settings::load_toml(&path).unwrap().is_none());
+
+        // Simulate what persist_selected_model now does for the Ok(None) case.
+        let settings = Settings {
+            selected_model: Some("new-model".to_string()),
+            ..Default::default()
+        };
+        settings.save_toml(&path).unwrap();
+
+        // Verify the model survived.
+        let loaded = Settings::load_toml(&path).unwrap().unwrap();
+        assert_eq!(loaded.selected_model, Some("new-model".to_string()));
+    }
+
     #[test]
     fn toml_missing_file_returns_none() {
         let result = Settings::load_toml(std::path::Path::new("/tmp/nonexistent_config.toml"));

src/testing/mod.rs

@@ -563,6 +563,7 @@ impl TestHarnessBuilder {
             document_extraction: None,
             sandbox_readiness: crate::agent::routine_engine::SandboxReadiness::DisabledByConfig,
             builder: None,
+            llm_backend: "nearai".to_string(),
         };
 
         TestHarness {

src/tools/builtin/extension_tools.rs

@@ -130,7 +130,7 @@ impl Tool for ToolInstallTool {
     async fn execute(
         &self,
         params: serde_json::Value,
-        _ctx: &JobContext,
+        ctx: &JobContext,
     ) -> Result<ToolOutput, ToolError> {
         let start = std::time::Instant::now();
 
@@ -150,7 +150,7 @@ impl Tool for ToolInstallTool {
 
         let result = self
             .manager
-            .install(name, url, kind_hint)
+            .install(name, url, kind_hint, &ctx.user_id)
             .await
             .map_err(|e| ToolError::ExecutionFailed(e.to_string()))?;
 
@@ -205,7 +205,7 @@ impl Tool for ToolAuthTool {
     async fn execute(
         &self,
         params: serde_json::Value,
-        _ctx: &JobContext,
+        ctx: &JobContext,
     ) -> Result<ToolOutput, ToolError> {
         let start = std::time::Instant::now();
 
@@ -213,13 +213,13 @@ impl Tool for ToolAuthTool {
 
         let result = self
             .manager
-            .auth(name)
+            .auth(name, &ctx.user_id)
             .await
             .map_err(|e| ToolError::ExecutionFailed(e.to_string()))?;
 
         // Auto-activate after successful auth so tools are available immediately
         if result.is_authenticated() {
-            match self.manager.activate(name).await {
+            match self.manager.activate(name, &ctx.user_id).await {
                 Ok(activate_result) => {
                     let output = serde_json::json!({
                         "status": "authenticated_and_activated",
@@ -304,13 +304,13 @@ impl Tool for ToolActivateTool {
     async fn execute(
         &self,
         params: serde_json::Value,
-        _ctx: &JobContext,
+        ctx: &JobContext,
     ) -> Result<ToolOutput, ToolError> {
         let start = std::time::Instant::now();
 
         let name = require_str(&params, "name")?;
 
-        match self.manager.activate(name).await {
+        match self.manager.activate(name, &ctx.user_id).await {
             Ok(result) => {
                 let output = serde_json::to_value(&result)
                     .unwrap_or_else(|_| serde_json::json!({"error": "serialization failed"}));
@@ -329,12 +329,12 @@ impl Tool for ToolActivateTool {
 
                 // Activation failed due to missing auth; initiate auth flow
                 // so the agent loop can show the auth card.
-                match self.manager.auth(name).await {
+                match self.manager.auth(name, &ctx.user_id).await {
                     Ok(auth_result) if auth_result.is_authenticated() => {
                         // Auth succeeded (e.g. env var was set); retry activation.
                         let result = self
                             .manager
-                            .activate(name)
+                            .activate(name, &ctx.user_id)
                             .await
                             .map_err(|e| ToolError::ExecutionFailed(e.to_string()))?;
                         let output = serde_json::to_value(&result).unwrap_or_else(
@@ -404,7 +404,7 @@ impl Tool for ToolListTool {
     async fn execute(
         &self,
         params: serde_json::Value,
-        _ctx: &JobContext,
+        ctx: &JobContext,
     ) -> Result<ToolOutput, ToolError> {
         let start = std::time::Instant::now();
 
@@ -425,7 +425,7 @@ impl Tool for ToolListTool {
 
         let extensions = self
             .manager
-            .list(kind_filter, include_available)
+            .list(kind_filter, include_available, &ctx.user_id)
             .await
             .map_err(|e| ToolError::ExecutionFailed(e.to_string()))?;
 
@@ -477,7 +477,7 @@ impl Tool for ToolRemoveTool {
     async fn execute(
         &self,
         params: serde_json::Value,
-        _ctx: &JobContext,
+        ctx: &JobContext,
     ) -> Result<ToolOutput, ToolError> {
         let start = std::time::Instant::now();
 
@@ -485,7 +485,7 @@ impl Tool for ToolRemoveTool {
 
         let message = self
             .manager
-            .remove(name)
+            .remove(name, &ctx.user_id)
             .await
             .map_err(|e| ToolError::ExecutionFailed(e.to_string()))?;
 
@@ -541,7 +541,7 @@ impl Tool for ToolUpgradeTool {
     async fn execute(
         &self,
         params: serde_json::Value,
-        _ctx: &JobContext,
+        ctx: &JobContext,
     ) -> Result<ToolOutput, ToolError> {
         let start = std::time::Instant::now();
 
@@ -549,7 +549,7 @@ impl Tool for ToolUpgradeTool {
 
         let result = self
             .manager
-            .upgrade(name)
+            .upgrade(name, &ctx.user_id)
             .await
             .map_err(|e| ToolError::ExecutionFailed(e.to_string()))?;
 
@@ -603,7 +603,7 @@ impl Tool for ExtensionInfoTool {
     async fn execute(
         &self,
         params: serde_json::Value,
-        _ctx: &JobContext,
+        ctx: &JobContext,
     ) -> Result<ToolOutput, ToolError> {
         let start = std::time::Instant::now();
 
@@ -611,7 +611,7 @@ impl Tool for ExtensionInfoTool {
 
         let info = self
             .manager
-            .extension_info(name)
+            .extension_info(name, &ctx.user_id)
             .await
             .map_err(|e| ToolError::ExecutionFailed(e.to_string()))?;
 

src/tools/builtin/job.rs

@@ -17,7 +17,6 @@ use uuid::Uuid;
 
 use crate::bootstrap::ironclaw_base_dir;
 use crate::channels::IncomingMessage;
-use crate::channels::web::types::SseEvent;
 use crate::context::{ContextManager, JobContext, JobState};
 use crate::db::Database;
 use crate::history::SandboxJobRecord;
@@ -25,6 +24,7 @@ use crate::orchestrator::auth::CredentialGrant;
 use crate::orchestrator::job_manager::{ContainerJobManager, JobMode};
 use crate::secrets::SecretsStore;
 use crate::tools::tool::{ApprovalRequirement, Tool, ToolError, ToolOutput, require_str};
+use ironclaw_common::AppEvent;
 
 /// Lazy scheduler reference, filled after Agent::new creates the Scheduler.
 ///
@@ -85,7 +85,7 @@ pub struct CreateJobTool {
     job_manager: Option<Arc<ContainerJobManager>>,
     store: Option<Arc<dyn Database>>,
     /// Broadcast sender for job events (used to subscribe a monitor).
-    event_tx: Option<tokio::sync::broadcast::Sender<(Uuid, SseEvent)>>,
+    event_tx: Option<tokio::sync::broadcast::Sender<(Uuid, String, AppEvent)>>,
     /// Injection channel for pushing messages into the agent loop.
     inject_tx: Option<tokio::sync::mpsc::Sender<IncomingMessage>>,
     /// Encrypted secrets store for validating credential grants.
@@ -120,7 +120,7 @@ impl CreateJobTool {
     /// monitor that forwards Claude Code output to the main agent loop.
     pub fn with_monitor_deps(
         mut self,
-        event_tx: tokio::sync::broadcast::Sender<(Uuid, SseEvent)>,
+        event_tx: tokio::sync::broadcast::Sender<(Uuid, String, AppEvent)>,
         inject_tx: tokio::sync::mpsc::Sender<IncomingMessage>,
     ) -> Self {
         self.event_tx = Some(event_tx);

src/tools/builtin/memory.rs

@@ -21,6 +21,35 @@ use crate::context::JobContext;
 use crate::tools::tool::{Tool, ToolError, ToolOutput, require_str};
 use crate::workspace::{Workspace, paths};
 
+// ── WorkspaceResolver ──────────────────────────────────────────────
+
+/// Resolves a workspace for a given user ID.
+///
+/// In single-user mode, always returns the same workspace.
+/// In multi-tenant mode, creates per-user workspaces on demand.
+#[async_trait]
+pub trait WorkspaceResolver: Send + Sync {
+    async fn resolve(&self, user_id: &str) -> Arc<Workspace>;
+}
+
+/// Returns a fixed workspace regardless of user ID (single-user mode).
+pub struct FixedWorkspaceResolver {
+    workspace: Arc<Workspace>,
+}
+
+impl FixedWorkspaceResolver {
+    pub fn new(workspace: Arc<Workspace>) -> Self {
+        Self { workspace }
+    }
+}
+
+#[async_trait]
+impl WorkspaceResolver for FixedWorkspaceResolver {
+    async fn resolve(&self, _user_id: &str) -> Arc<Workspace> {
+        Arc::clone(&self.workspace)
+    }
+}
+
 /// Detect paths that are clearly local filesystem references, not workspace-memory docs.
 ///
 /// Examples:
@@ -62,13 +91,20 @@ fn map_write_err(e: crate::error::WorkspaceError) -> ToolError {
 /// The agent should call this tool before answering questions about
 /// prior work, decisions, preferences, or any historical context.
 pub struct MemorySearchTool {
-    workspace: Arc<Workspace>,
+    resolver: Arc<dyn WorkspaceResolver>,
 }
 
 impl MemorySearchTool {
-    /// Create a new memory search tool.
+    /// Create a new memory search tool with a workspace resolver.
-    pub fn new(workspace: Arc<Workspace>) -> Self {
+    pub fn new(resolver: Arc<dyn WorkspaceResolver>) -> Self {
-        Self { workspace }
+        Self { resolver }
+    }
+
+    /// Create from a fixed workspace (backward compatibility).
+    pub fn from_workspace(workspace: Arc<Workspace>) -> Self {
+        Self {
+            resolver: Arc::new(FixedWorkspaceResolver::new(workspace)),
+        }
     }
 }
 
@@ -107,7 +143,7 @@ impl Tool for MemorySearchTool {
     async fn execute(
         &self,
         params: serde_json::Value,
-        _ctx: &JobContext,
+        ctx: &JobContext,
     ) -> Result<ToolOutput, ToolError> {
         let start = std::time::Instant::now();
 
@@ -119,8 +155,8 @@ impl Tool for MemorySearchTool {
             .unwrap_or(5)
             .min(20) as usize;
 
-        let results = self
+        let workspace = self.resolver.resolve(&ctx.user_id).await;
-            .workspace
+        let results = workspace
             .search(query, limit)
             .await
             .map_err(|e| ToolError::ExecutionFailed(format!("Search failed: {}", e)))?;
@@ -151,13 +187,20 @@ impl Tool for MemorySearchTool {
 /// Use this to persist important information that should be remembered
 /// across sessions: decisions, preferences, facts, lessons learned.
 pub struct MemoryWriteTool {
-    workspace: Arc<Workspace>,
+    resolver: Arc<dyn WorkspaceResolver>,
 }
 
 impl MemoryWriteTool {
-    /// Create a new memory write tool.
+    /// Create a new memory write tool with a workspace resolver.
-    pub fn new(workspace: Arc<Workspace>) -> Self {
+    pub fn new(resolver: Arc<dyn WorkspaceResolver>) -> Self {
-        Self { workspace }
+        Self { resolver }
+    }
+
+    /// Create from a fixed workspace (backward compatibility).
+    pub fn from_workspace(workspace: Arc<Workspace>) -> Self {
+        Self {
+            resolver: Arc::new(FixedWorkspaceResolver::new(workspace)),
+        }
     }
 }
 
@@ -231,19 +274,21 @@ impl Tool for MemoryWriteTool {
             )));
         }
 
+        let workspace = self.resolver.resolve(&ctx.user_id).await;
+
         // Bootstrap target: clear BOOTSTRAP.md to mark first-run ritual complete.
         // Handled early because it accepts empty content (unlike other targets).
         if target == "bootstrap" {
             // Write empty content to effectively disable the bootstrap injection.
             // system_prompt_for_context() skips empty files.
-            self.workspace
+            workspace
                 .write(paths::BOOTSTRAP, "")
                 .await
                 .map_err(map_write_err)?;
 
             // Also set the in-memory flag so BOOTSTRAP.md injection stops
             // immediately without waiting for a restart.
-            self.workspace.mark_bootstrap_completed();
+            workspace.mark_bootstrap_completed();
 
             let output = serde_json::json!({
                 "status": "cleared",
@@ -289,12 +334,12 @@ impl Tool for MemoryWriteTool {
         // Otherwise, use default workspace methods (which include injection scanning).
         let layer_result = if let Some(layer_name) = layer {
             let result = if append {
-                self.workspace
+                workspace
                     .append_to_layer(layer_name, &resolved_path, content, force)
                     .await
                     .map_err(map_write_err)?
             } else {
-                self.workspace
+                workspace
                     .write_to_layer(layer_name, &resolved_path, content, force)
                     .await
                     .map_err(map_write_err)?
@@ -307,31 +352,33 @@ impl Tool for MemoryWriteTool {
             match target {
                 "memory" => {
                     if append {
-                        self.workspace
+                        workspace
                             .append_memory(content)
                             .await
                             .map_err(map_write_err)?;
                     } else {
-                        self.workspace
+                        workspace
                             .write(paths::MEMORY, content)
                             .await
                             .map_err(map_write_err)?;
                     }
                 }
                 "daily_log" => {
-                    self.workspace
+                    let tz = crate::timezone::parse_timezone(&ctx.user_timezone)
+                        .unwrap_or(chrono_tz::Tz::UTC);
+                    workspace
                         .append_daily_log_tz(content, tz)
                         .await
                         .map_err(map_write_err)?;
                 }
                 _ => {
                     if append {
-                        self.workspace
+                        workspace
                             .append(&resolved_path, content)
                             .await
                             .map_err(map_write_err)?;
                     } else {
-                        self.workspace
+                        workspace
                             .write(&resolved_path, content)
                             .await
                             .map_err(map_write_err)?;
@@ -361,12 +408,12 @@ impl Tool for MemoryWriteTool {
         };
         let mut synced_docs: Vec<&str> = Vec::new();
         if normalized_path == paths::PROFILE {
-            match self.workspace.sync_profile_documents().await {
+            match workspace.sync_profile_documents().await {
                 Ok(true) => {
                     tracing::info!("profile write: synced USER.md + assistant-directives.md");
                     synced_docs.extend_from_slice(&[paths::USER, paths::ASSISTANT_DIRECTIVES]);
 
-                    self.workspace.mark_bootstrap_completed();
+                    workspace.mark_bootstrap_completed();
                     let toml_path = crate::settings::Settings::default_toml_path();
                     if let Ok(Some(mut settings)) = crate::settings::Settings::load_toml(&toml_path)
                         && !settings.profile_onboarding_completed
@@ -416,13 +463,20 @@ impl Tool for MemoryWriteTool {
 ///
 /// Use this to read the full content of any file in the workspace.
 pub struct MemoryReadTool {
-    workspace: Arc<Workspace>,
+    resolver: Arc<dyn WorkspaceResolver>,
 }
 
 impl MemoryReadTool {
-    /// Create a new memory read tool.
+    /// Create a new memory read tool with a workspace resolver.
-    pub fn new(workspace: Arc<Workspace>) -> Self {
+    pub fn new(resolver: Arc<dyn WorkspaceResolver>) -> Self {
-        Self { workspace }
+        Self { resolver }
+    }
+
+    /// Create from a fixed workspace (backward compatibility).
+    pub fn from_workspace(workspace: Arc<Workspace>) -> Self {
+        Self {
+            resolver: Arc::new(FixedWorkspaceResolver::new(workspace)),
+        }
     }
 }
 
@@ -456,7 +510,7 @@ impl Tool for MemoryReadTool {
     async fn execute(
         &self,
         params: serde_json::Value,
-        _ctx: &JobContext,
+        ctx: &JobContext,
     ) -> Result<ToolOutput, ToolError> {
         let start = std::time::Instant::now();
 
@@ -470,8 +524,8 @@ impl Tool for MemoryReadTool {
             )));
         }
 
-        let doc = self
+        let workspace = self.resolver.resolve(&ctx.user_id).await;
-            .workspace
+        let doc = workspace
             .read(path)
             .await
             .map_err(|e| ToolError::ExecutionFailed(format!("Read failed: {}", e)))?;
@@ -495,20 +549,27 @@ impl Tool for MemoryReadTool {
 ///
 /// Returns a hierarchical view of files and directories with configurable depth.
 pub struct MemoryTreeTool {
-    workspace: Arc<Workspace>,
+    resolver: Arc<dyn WorkspaceResolver>,
 }
 
 impl MemoryTreeTool {
-    /// Create a new memory tree tool.
+    /// Create a new memory tree tool with a workspace resolver.
-    pub fn new(workspace: Arc<Workspace>) -> Self {
+    pub fn new(resolver: Arc<dyn WorkspaceResolver>) -> Self {
-        Self { workspace }
+        Self { resolver }
+    }
+
+    /// Create from a fixed workspace (backward compatibility).
+    pub fn from_workspace(workspace: Arc<Workspace>) -> Self {
+        Self {
+            resolver: Arc::new(FixedWorkspaceResolver::new(workspace)),
+        }
     }
 
     /// Recursively build tree structure.
     ///
     /// Returns a compact format where directories end with `/` and may have children.
     async fn build_tree(
-        &self,
+        workspace: &Arc<Workspace>,
         path: &str,
         current_depth: usize,
         max_depth: usize,
@@ -517,8 +578,7 @@ impl MemoryTreeTool {
             return Ok(Vec::new());
         }
 
-        let entries = self
+        let entries = workspace
-            .workspace
             .list(path)
             .await
             .map_err(|e| ToolError::ExecutionFailed(format!("Tree failed: {}", e)))?;
@@ -533,8 +593,13 @@ impl MemoryTreeTool {
             };
 
             if entry.is_directory && current_depth < max_depth {
-                let children =
+                let children = Box::pin(Self::build_tree(
-                    Box::pin(self.build_tree(&entry.path, current_depth + 1, max_depth)).await?;
+                    workspace,
+                    &entry.path,
+                    current_depth + 1,
+                    max_depth,
+                ))
+                .await?;
                 if children.is_empty() {
                     result.push(serde_json::Value::String(display_path));
                 } else {
@@ -584,7 +649,7 @@ impl Tool for MemoryTreeTool {
     async fn execute(
         &self,
         params: serde_json::Value,
-        _ctx: &JobContext,
+        ctx: &JobContext,
     ) -> Result<ToolOutput, ToolError> {
         let start = std::time::Instant::now();
 
@@ -596,7 +661,8 @@ impl Tool for MemoryTreeTool {
             .unwrap_or(1)
             .clamp(1, 10) as usize;
 
-        let tree = self.build_tree(path, 1, depth).await?;
+        let workspace = self.resolver.resolve(&ctx.user_id).await;
+        let tree = Self::build_tree(&workspace, path, 1, depth).await?;
 
         // Compact output: just the tree array
         Ok(ToolOutput::success(
@@ -650,7 +716,7 @@ mod tests {
         #[test]
         fn test_memory_search_schema() {
             let workspace = make_test_workspace();
-            let tool = MemorySearchTool::new(workspace);
+            let tool = MemorySearchTool::from_workspace(workspace);
 
             assert_eq!(tool.name(), "memory_search");
             assert!(!tool.requires_sanitization());
@@ -668,7 +734,7 @@ mod tests {
         #[test]
         fn test_memory_write_schema() {
             let workspace = make_test_workspace();
-            let tool = MemoryWriteTool::new(workspace);
+            let tool = MemoryWriteTool::from_workspace(workspace);
 
             assert_eq!(tool.name(), "memory_write");
 
@@ -681,7 +747,7 @@ mod tests {
         #[test]
         fn test_memory_read_schema() {
             let workspace = make_test_workspace();
-            let tool = MemoryReadTool::new(workspace);
+            let tool = MemoryReadTool::from_workspace(workspace);
 
             assert_eq!(tool.name(), "memory_read");
 
@@ -698,7 +764,7 @@ mod tests {
         #[test]
         fn test_memory_tree_schema() {
             let workspace = make_test_workspace();
-            let tool = MemoryTreeTool::new(workspace);
+            let tool = MemoryTreeTool::from_workspace(workspace);
 
             assert_eq!(tool.name(), "memory_tree");
 
@@ -711,7 +777,7 @@ mod tests {
         #[tokio::test]
         async fn test_memory_write_rejects_injection_to_identity_file() {
             let workspace = make_test_workspace();
-            let tool = MemoryWriteTool::new(workspace);
+            let tool = MemoryWriteTool::from_workspace(workspace);
             let ctx = JobContext::default();
 
             let params = serde_json::json!({
@@ -733,4 +799,176 @@ mod tests {
             }
         }
     }
+
+    // Regression tests for per-user workspace scoping (multi-tenant mode).
+    // See: https://github.com/nearai/ironclaw/pull/1118
+    // Bug: memory tools used a single startup workspace regardless of which
+    // user was chatting. Fix: resolve workspace per-request via JobContext.user_id.
+
+    #[cfg(feature = "postgres")]
+    mod resolver_tests {
+        use super::*;
+
+        fn make_test_workspace_for_user(user_id: &str) -> Arc<Workspace> {
+            Arc::new(Workspace::new(
+                user_id,
+                deadpool_postgres::Pool::builder(deadpool_postgres::Manager::new(
+                    tokio_postgres::Config::new(),
+                    tokio_postgres::NoTls,
+                ))
+                .build()
+                .unwrap(),
+            ))
+        }
+
+        #[tokio::test]
+        async fn test_fixed_workspace_resolver_ignores_user_id() {
+            let ws = make_test_workspace_for_user("alice");
+            let resolver = FixedWorkspaceResolver::new(Arc::clone(&ws));
+
+            let ws_alice = resolver.resolve("alice").await;
+            let ws_bob = resolver.resolve("bob").await;
+
+            // Both should return the exact same Arc (pointer equality)
+            assert!(Arc::ptr_eq(&ws_alice, &ws_bob));
+            assert_eq!(ws_alice.user_id(), "alice");
+        }
+
+        /// Tracking resolver that records which user_ids were requested.
+        struct TrackingWorkspaceResolver {
+            inner: FixedWorkspaceResolver,
+            resolved_users: std::sync::Mutex<Vec<String>>,
+        }
+
+        impl TrackingWorkspaceResolver {
+            fn new(workspace: Arc<Workspace>) -> Self {
+                Self {
+                    inner: FixedWorkspaceResolver::new(workspace),
+                    resolved_users: std::sync::Mutex::new(Vec::new()),
+                }
+            }
+
+            fn resolved_users(&self) -> Vec<String> {
+                self.resolved_users.lock().unwrap().clone()
+            }
+        }
+
+        #[async_trait]
+        impl WorkspaceResolver for TrackingWorkspaceResolver {
+            async fn resolve(&self, user_id: &str) -> Arc<Workspace> {
+                self.resolved_users
+                    .lock()
+                    .unwrap()
+                    .push(user_id.to_string());
+                self.inner.resolve(user_id).await
+            }
+        }
+
+        #[tokio::test]
+        async fn test_memory_search_uses_job_context_user_id() {
+            let ws = make_test_workspace_for_user("default");
+            let tracker = Arc::new(TrackingWorkspaceResolver::new(ws));
+            let tool = MemorySearchTool::new(tracker.clone() as Arc<dyn WorkspaceResolver>);
+
+            // Execute with user_id "alice"
+            let ctx_alice = JobContext::with_user("alice", "test", "test");
+            let params = serde_json::json!({"query": "test"});
+            // The search will fail (no real DB) but we only care about resolver call
+            let _ = tool.execute(params, &ctx_alice).await;
+
+            // Execute with user_id "bob"
+            let ctx_bob = JobContext::with_user("bob", "test", "test");
+            let params = serde_json::json!({"query": "test"});
+            let _ = tool.execute(params, &ctx_bob).await;
+
+            let resolved = tracker.resolved_users();
+            assert_eq!(resolved, vec!["alice", "bob"]);
+        }
+
+        #[tokio::test]
+        async fn test_memory_write_uses_job_context_user_id() {
+            let ws = make_test_workspace_for_user("default");
+            let tracker = Arc::new(TrackingWorkspaceResolver::new(ws));
+            let tool = MemoryWriteTool::new(tracker.clone() as Arc<dyn WorkspaceResolver>);
+
+            // Execute with user_id "alice"
+            let ctx_alice = JobContext::with_user("alice", "test", "test");
+            let params = serde_json::json!({
+                "content": "remember this",
+                "target": "daily_log",
+            });
+            let _ = tool.execute(params, &ctx_alice).await;
+
+            // Execute with user_id "bob"
+            let ctx_bob = JobContext::with_user("bob", "test", "test");
+            let params = serde_json::json!({
+                "content": "remember that",
+                "target": "daily_log",
+            });
+            let _ = tool.execute(params, &ctx_bob).await;
+
+            let resolved = tracker.resolved_users();
+            assert_eq!(resolved, vec!["alice", "bob"]);
+        }
+    }
+
+    #[cfg(feature = "libsql")]
+    mod per_user_resolver_tests {
+        use super::*;
+
+        async fn make_test_db() -> Arc<dyn crate::db::Database> {
+            use crate::db::libsql::LibSqlBackend;
+            let temp_dir = tempfile::tempdir().expect("tempdir");
+            let db_path = temp_dir.path().join("resolver_test.db");
+            let backend = LibSqlBackend::new_local(&db_path)
+                .await
+                .expect("LibSqlBackend");
+            <LibSqlBackend as crate::db::Database>::run_migrations(&backend)
+                .await
+                .expect("migrations");
+            // Leak the tempdir so it outlives the test (cleaned up on process exit).
+            std::mem::forget(temp_dir);
+            Arc::new(backend)
+        }
+
+        #[tokio::test]
+        async fn test_workspace_pool_resolver_returns_different_workspaces() {
+            let db = make_test_db().await;
+
+            let pool = crate::channels::web::server::WorkspacePool::new(
+                db,
+                None,
+                crate::workspace::EmbeddingCacheConfig::default(),
+                crate::config::WorkspaceSearchConfig::default(),
+                crate::config::WorkspaceConfig::default(),
+            );
+
+            let ws_alice = pool.resolve("alice").await;
+            let ws_bob = pool.resolve("bob").await;
+
+            // Different user IDs should get different workspaces
+            assert_eq!(ws_alice.user_id(), "alice");
+            assert_eq!(ws_bob.user_id(), "bob");
+            assert!(!Arc::ptr_eq(&ws_alice, &ws_bob));
+        }
+
+        #[tokio::test]
+        async fn test_workspace_pool_resolver_caches_workspace() {
+            let db = make_test_db().await;
+
+            let pool = crate::channels::web::server::WorkspacePool::new(
+                db,
+                None,
+                crate::workspace::EmbeddingCacheConfig::default(),
+                crate::config::WorkspaceSearchConfig::default(),
+                crate::config::WorkspaceConfig::default(),
+            );
+
+            let ws1 = pool.resolve("alice").await;
+            let ws2 = pool.resolve("alice").await;
+
+            // Same user_id should return the same cached Arc (pointer equality)
+            assert!(Arc::ptr_eq(&ws1, &ws2));
+        }
+    }
 }

src/tools/builtin/mod.rs

@@ -6,7 +6,7 @@ mod file;
 mod http;
 mod job;
 mod json;
-mod memory;
+pub mod memory;
 mod message;
 pub mod path_utils;
 mod restart;

src/tools/builtin/routine.rs

@@ -140,7 +140,8 @@ fn execution_properties() -> Value {
         },
         "use_tools": {
             "type": "boolean",
-            "description": "Only applies to lightweight mode. When true, safe non-approval tools are available."
+            "default": true,
+            "description": "Only applies to lightweight mode. New lightweight routines default this to true; when enabled, the routine can use the owner's live autonomous tool scope."
         },
         "max_tool_rounds": {
             "type": "integer",
@@ -290,7 +291,7 @@ fn routine_request_discovery_schema() -> Value {
 fn lightweight_execution_variant() -> Value {
     serde_json::json!({
         "type": "object",
-        "description": "Default lightweight execution. Applies when execution is omitted or execution.mode='lightweight'.",
+        "description": "Default lightweight execution. Applies when execution is omitted or execution.mode='lightweight'. New lightweight routines default to tools enabled unless execution.use_tools=false is set.",
         "properties": {
             "mode": {
                 "type": "string",
@@ -304,7 +305,8 @@ fn lightweight_execution_variant() -> Value {
             },
             "use_tools": {
                 "type": "boolean",
-                "description": "When true, safe non-approval tools are available."
+                "default": true,
+                "description": "Defaults to true for new lightweight routines. When enabled, the routine can use the owner's live autonomous tool scope."
             },
             "max_tool_rounds": {
                 "type": "integer",
@@ -335,7 +337,7 @@ fn full_job_execution_variant() -> Value {
 fn execution_discovery_schema() -> Value {
     serde_json::json!({
         "type": "object",
-        "description": "Optional execution settings. Omit this block for the default lightweight mode.",
+        "description": "Optional execution settings. Omit this block for the default lightweight mode with tools enabled.",
         "properties": execution_properties(),
         "oneOf": [
             lightweight_execution_variant(),
@@ -408,7 +410,8 @@ fn routine_create_tool_summary() -> ToolDiscoverySummary {
             "execution.mode='full_job' uses the owner's live autonomous tool scope and ignores use_tools, max_tool_rounds, and context_paths.".into(),
         ],
         notes: vec![
-            "Omitting execution defaults to lightweight mode.".into(),
+            "Omitting execution defaults to lightweight mode with tools enabled.".into(),
+            "Set execution.use_tools=false to keep a new lightweight routine text-only.".into(),
             "Omitting delivery.user falls back to the owner's last-seen notification target.".into(),
             "advanced.cooldown_secs defaults to 300.".into(),
             "Legacy flat aliases are still accepted for compatibility, but grouped fields are preferred.".into(),
@@ -605,7 +608,8 @@ fn routine_create_schema(include_compatibility_aliases: bool) -> Value {
 }
 
 pub(crate) fn routine_create_parameters_schema() -> Value {
-    routine_create_schema(false)
+    static CACHE: OnceLock<Value> = OnceLock::new();
+    CACHE.get_or_init(|| routine_create_schema(false)).clone()
 }
 
 fn routine_create_discovery_schema() -> Value {
@@ -852,11 +856,15 @@ fn parse_execution_mode(value: Option<String>) -> Result<NormalizedExecutionMode
     }
 }
 
-fn parse_routine_execution(params: &Value) -> Result<NormalizedExecutionRequest, ToolError> {
+fn parse_routine_execution(
+    params: &Value,
+    default_use_tools: bool,
+) -> Result<NormalizedExecutionRequest, ToolError> {
     let mode = parse_execution_mode(string_field(params, "execution", "mode", &["action_type"]))?;
     let context_paths =
         string_array_field(params, "execution", "context_paths", &["context_paths"]);
-    let use_tools = bool_field(params, "execution", "use_tools", &["use_tools"]).unwrap_or(false);
+    let use_tools =
+        bool_field(params, "execution", "use_tools", &["use_tools"]).unwrap_or(default_use_tools);
     let max_tool_rounds = u64_field(params, "execution", "max_tool_rounds", &["max_tool_rounds"])
         .unwrap_or(3)
         .clamp(1, crate::agent::routine::MAX_TOOL_ROUNDS_LIMIT as u64)
@@ -888,7 +896,7 @@ fn parse_routine_create_request(
         .unwrap_or("")
         .to_string();
     let trigger = parse_routine_trigger(params)?;
-    let execution = parse_routine_execution(params)?;
+    let execution = parse_routine_execution(params, true)?;
     let delivery = parse_routine_delivery(params);
     let cooldown_secs =
         u64_field(params, "advanced", "cooldown_secs", &["cooldown_secs"]).unwrap_or(300);
@@ -1007,7 +1015,8 @@ fn event_emit_schema(include_source_alias: bool) -> Value {
 }
 
 pub(crate) fn event_emit_parameters_schema() -> Value {
-    event_emit_schema(false)
+    static CACHE: OnceLock<Value> = OnceLock::new();
+    CACHE.get_or_init(|| event_emit_schema(false)).clone()
 }
 
 fn event_emit_discovery_schema() -> Value {
@@ -1863,6 +1872,56 @@ mod tests {
         );
     }
 
+    #[test]
+    fn parses_lightweight_create_with_tools_enabled_by_default() {
+        let params = serde_json::json!({
+            "name": "manual-check",
+            "prompt": "Inspect the repo for issues.",
+            "request": {
+                "kind": "manual"
+            }
+        });
+
+        let parsed = parse_routine_create_request(&params).expect("parse default lightweight");
+
+        assert!(
+            matches!(parsed.execution.mode, NormalizedExecutionMode::Lightweight),
+            "expected lightweight execution mode",
+        );
+        assert!(
+            parsed.execution.use_tools,
+            "new lightweight routines should default use_tools=true",
+        );
+        assert_eq!(parsed.execution.max_tool_rounds, 3);
+    }
+
+    #[test]
+    fn parses_lightweight_create_with_explicit_tools_disabled() {
+        let params = serde_json::json!({
+            "name": "manual-check",
+            "prompt": "Inspect the repo for issues.",
+            "request": {
+                "kind": "manual"
+            },
+            "execution": {
+                "use_tools": false
+            }
+        });
+
+        let parsed =
+            parse_routine_create_request(&params).expect("parse lightweight with tools disabled");
+
+        assert!(
+            matches!(parsed.execution.mode, NormalizedExecutionMode::Lightweight),
+            "expected lightweight execution mode",
+        );
+        assert!(
+            !parsed.execution.use_tools,
+            "explicit use_tools=false should be preserved",
+        );
+        assert_eq!(parsed.execution.max_tool_rounds, 3);
+    }
+
     #[test]
     fn parses_context_paths_with_trim_drop_empty_and_stable_dedupe() {
         let params = serde_json::json!({
@@ -2201,6 +2260,20 @@ mod tests {
                 .any(|rule| rule.contains("request.kind='cron'")),
             "summary should explain cron requirement",
         );
+        assert!(
+            summary
+                .notes
+                .iter()
+                .any(|note| note.contains("lightweight mode with tools enabled")),
+            "summary should mention the new lightweight default",
+        );
+        assert!(
+            summary
+                .notes
+                .iter()
+                .any(|note| note.contains("execution.use_tools=false")),
+            "summary should mention the text-only opt-out",
+        );
         assert!(
             summary
                 .notes

src/tools/registry.rs

@@ -334,15 +334,37 @@ impl ToolRegistry {
         tracing::debug!("Registered 5 development tools");
     }
 
-    /// Register memory tools with a workspace.
+    /// Register memory tools with a workspace resolver.
+    ///
+    /// Memory tools require a workspace resolver for persistence. Call this after
+    /// `register_builtin_tools()` if you have a workspace available.
+    pub fn register_memory_tools_with_resolver(
+        &self,
+        resolver: Arc<dyn crate::tools::builtin::memory::WorkspaceResolver>,
+    ) {
+        self.register_sync(Arc::new(MemorySearchTool::new(Arc::clone(&resolver))));
+        self.register_sync(Arc::new(MemoryWriteTool::new(Arc::clone(&resolver))));
+        self.register_sync(Arc::new(MemoryReadTool::new(Arc::clone(&resolver))));
+        self.register_sync(Arc::new(MemoryTreeTool::new(resolver)));
+
+        tracing::debug!("Registered 4 memory tools");
+    }
+
+    /// Register memory tools with a fixed workspace (backward compatibility).
     ///
     /// Memory tools require a workspace for persistence. Call this after
     /// `register_builtin_tools()` if you have a workspace available.
     pub fn register_memory_tools(&self, workspace: Arc<Workspace>) {
-        self.register_sync(Arc::new(MemorySearchTool::new(Arc::clone(&workspace))));
+        self.register_sync(Arc::new(MemorySearchTool::from_workspace(Arc::clone(
-        self.register_sync(Arc::new(MemoryWriteTool::new(Arc::clone(&workspace))));
+            &workspace,
-        self.register_sync(Arc::new(MemoryReadTool::new(Arc::clone(&workspace))));
+        ))));
-        self.register_sync(Arc::new(MemoryTreeTool::new(workspace)));
+        self.register_sync(Arc::new(MemoryWriteTool::from_workspace(Arc::clone(
+            &workspace,
+        ))));
+        self.register_sync(Arc::new(MemoryReadTool::from_workspace(Arc::clone(
+            &workspace,
+        ))));
+        self.register_sync(Arc::new(MemoryTreeTool::from_workspace(workspace)));
 
         tracing::debug!("Registered 4 memory tools");
     }
@@ -361,7 +383,7 @@ impl ToolRegistry {
         job_manager: Option<Arc<ContainerJobManager>>,
         store: Option<Arc<dyn Database>>,
         job_event_tx: Option<
-            tokio::sync::broadcast::Sender<(uuid::Uuid, crate::channels::web::types::SseEvent)>,
+            tokio::sync::broadcast::Sender<(uuid::Uuid, String, ironclaw_common::AppEvent)>,
         >,
         inject_tx: Option<tokio::sync::mpsc::Sender<crate::channels::IncomingMessage>>,
         prompt_queue: Option<PromptQueue>,

src/tools/wasm/capabilities_schema.rs

@@ -47,12 +47,6 @@ pub struct CapabilitiesFile {
     #[serde(default)]
     pub description: Option<String>,
 
-    /// JSON Schema for the tool's input parameters.
-    /// Used as the `Tool::parameters_schema()` return value.
-    /// If omitted, a permissive fallback is used (with a warning).
-    #[serde(default)]
-    pub parameters: Option<serde_json::Value>,
-
     /// Extension version (semver).
     #[serde(default)]
     pub version: Option<String>,
@@ -103,9 +97,6 @@ pub struct CapabilitiesFile {
 
 /// Maximum length for the description field to prevent memory abuse.
 const MAX_DESCRIPTION_CHARS: usize = 4096;
-/// Maximum serialized size of the parameters schema JSON.
-const MAX_PARAMETERS_SCHEMA_BYTES: usize = 64 * 1024;
-
 impl CapabilitiesFile {
     /// Parse from JSON string.
     pub fn from_json(json: &str) -> Result<Self, serde_json::Error> {
@@ -135,18 +126,6 @@ impl CapabilitiesFile {
             );
             self.description = Some(truncated.to_string());
         }
-        // Drop oversized parameters schema (issue #977)
-        if let Some(ref params) = self.parameters {
-            let size = params.to_string().len();
-            if size > MAX_PARAMETERS_SCHEMA_BYTES {
-                tracing::warn!(
-                    "Capabilities parameters schema dropped ({} bytes exceeds {} limit)",
-                    size,
-                    MAX_PARAMETERS_SCHEMA_BYTES,
-                );
-                self.parameters = None;
-            }
-        }
     }
 
     /// Merge nested `capabilities` wrapper into top-level fields.
@@ -171,7 +150,6 @@ impl CapabilitiesFile {
         if let Some(inner) = self.capabilities.take() {
             let inner = inner.resolve_nested_inner(depth + 1);
             self.description = self.description.or(inner.description);
-            self.parameters = self.parameters.or(inner.parameters);
             self.http = self.http.or(inner.http);
             self.secrets = self.secrets.or(inner.secrets);
             self.tool_invoke = self.tool_invoke.or(inner.tool_invoke);
@@ -1424,26 +1402,12 @@ mod tests {
         );
     }
 
-    // ── Tool description and parameters schema ──────────────────────────
+    // ── Tool description ────────────────────────────────────────────────
 
     #[test]
-    fn test_parse_description_and_parameters() {
+    fn test_parse_description() {
         let json = r#"{
-            "description": "Search the web using Brave Search API",
+            "description": "Search the web using Brave Search API"
-            "parameters": {
-                "type": "object",
-                "properties": {
-                    "query": {
-                        "type": "string",
-                        "description": "Search query"
-                    },
-                    "count": {
-                        "type": "integer",
-                        "description": "Number of results"
-                    }
-                },
-                "required": ["query"]
-            }
         }"#;
 
         let caps = CapabilitiesFile::from_json(json).unwrap();
@@ -1451,28 +1415,10 @@ mod tests {
             caps.description.as_deref(),
             Some("Search the web using Brave Search API")
         );
-        let params = caps.parameters.unwrap();
-        assert_eq!(params["type"], "object");
-        assert!(params["properties"]["query"].is_object());
-        assert_eq!(params["required"][0], "query");
     }
 
     #[test]
-    fn test_parse_description_only() {
+    fn test_parse_without_description() {
-        let json = r#"{
-            "description": "A tool without explicit parameters schema"
-        }"#;
-
-        let caps = CapabilitiesFile::from_json(json).unwrap();
-        assert_eq!(
-            caps.description.as_deref(),
-            Some("A tool without explicit parameters schema")
-        );
-        assert!(caps.parameters.is_none());
-    }
-
-    #[test]
-    fn test_parse_without_description_or_parameters() {
         let json = r#"{
             "http": {
                 "allowlist": [{ "host": "api.example.com" }]
@@ -1484,24 +1430,28 @@ mod tests {
             caps.description.is_none(),
             "description should be None when not provided"
         );
-        assert!(
+    }
-            caps.parameters.is_none(),
+
-            "parameters should be None when not provided"
+    #[test]
-        );
+    fn test_parameters_field_silently_ignored() {
+        // Backward compat: old capabilities files with "parameters" still parse.
+        let json = r#"{
+            "description": "A tool",
+            "parameters": {
+                "type": "object",
+                "properties": { "action": { "type": "string" } }
+            }
+        }"#;
+
+        let caps = CapabilitiesFile::from_json(json).unwrap();
+        assert_eq!(caps.description.as_deref(), Some("A tool"));
     }
 
     #[test]
     fn test_resolve_nested_description_promoted() {
         let json = r#"{
             "capabilities": {
-                "description": "Inner tool description",
+                "description": "Inner tool description"
-                "parameters": {
-                    "type": "object",
-                    "properties": {
-                        "input": { "type": "string" }
-                    },
-                    "required": ["input"]
-                }
             }
         }"#;
 
@@ -1511,10 +1461,6 @@ mod tests {
             Some("Inner tool description"),
             "description should be promoted from inner capabilities"
         );
-        assert!(
-            caps.parameters.is_some(),
-            "parameters should be promoted from inner capabilities"
-        );
     }
 
     #[test]
@@ -1564,32 +1510,4 @@ mod tests {
             desc.len()
         );
     }
-
-    /// Regression test for issue #977: oversized parameters schema is dropped.
-    #[test]
-    fn test_oversized_parameters_schema_dropped() {
-        // Build a parameters schema larger than MAX_PARAMETERS_SCHEMA_BYTES
-        let mut properties = serde_json::Map::new();
-        for i in 0..2000 {
-            properties.insert(
-                format!("field_{i}"),
-                serde_json::json!({
-                    "type": "string",
-                    "description": "x".repeat(50)
-                }),
-            );
-        }
-        let schema = serde_json::json!({
-            "type": "object",
-            "properties": properties,
-        });
-        let json = serde_json::json!({
-            "parameters": schema,
-        });
-        let caps = CapabilitiesFile::from_json(&json.to_string()).unwrap();
-        assert!(
-            caps.parameters.is_none(),
-            "oversized parameters schema should be dropped"
-        );
-    }
 }

src/tools/wasm/loader.rs

@@ -123,73 +123,51 @@ impl WasmToolLoader {
         }
         let wasm_bytes = fs::read(wasm_path).await?;
 
-        // Read capabilities (optional) and extract OAuth refresh config,
+        // Read capabilities (optional) and extract OAuth refresh config
-        // tool description, and parameter schema.
+        // and tool description. Parameter schema is auto-derived from the
-        let (capabilities, oauth_refresh, description, schema) =
+        // WASM module's schema() export (see WasmToolSchemas::compact_schema).
-            if let Some(cap_path) = capabilities_path {
+        let (capabilities, oauth_refresh, description) = if let Some(cap_path) = capabilities_path {
-                if cap_path.exists() {
+            if cap_path.exists() {
-                    let cap_bytes = fs::read(cap_path).await?;
+                let cap_bytes = fs::read(cap_path).await?;
-                    let cap_file = CapabilitiesFile::from_bytes(&cap_bytes)
+                let cap_file = CapabilitiesFile::from_bytes(&cap_bytes)
-                        .map_err(|e| WasmLoadError::InvalidCapabilities(e.to_string()))?;
+                    .map_err(|e| WasmLoadError::InvalidCapabilities(e.to_string()))?;
-                    cap_file.validate(name);
+                cap_file.validate(name);
 
-                    // Check WIT version compatibility
+                // Check WIT version compatibility
-                    check_wit_version_compat(
+                check_wit_version_compat(
-                        name,
+                    name,
-                        cap_file.wit_version.as_deref(),
+                    cap_file.wit_version.as_deref(),
-                        crate::tools::wasm::WIT_TOOL_VERSION,
+                    crate::tools::wasm::WIT_TOOL_VERSION,
-                    )?;
+                )?;
 
-                    let caps = cap_file.to_capabilities();
+                let caps = cap_file.to_capabilities();
-                    let oauth = resolve_oauth_refresh_config(&cap_file);
+                let oauth = resolve_oauth_refresh_config(&cap_file);
-                    let desc = cap_file.description.clone();
+                let desc = cap_file.description.clone();
-                    // Validate parameters schema before accepting it.
+                if desc.is_none() {
-                    let params = cap_file.parameters.clone().and_then(|p| {
-                        let errors = crate::tools::validate_tool_schema(&p, name);
-                        if errors.is_empty() {
-                            Some(p)
-                        } else {
-                            tracing::warn!(
-                                tool = name,
-                                ?errors,
-                                "Invalid parameters schema in capabilities.json, \
-                                 using permissive fallback"
-                            );
-                            None
-                        }
-                    });
-                    if desc.is_none() {
-                        tracing::warn!(
-                            tool = name,
-                            path = %cap_path.display(),
-                            "Capabilities file missing \"description\" field; \
-                             tool will use generic fallback description"
-                        );
-                    }
-                    if params.is_none() && cap_file.parameters.is_none() {
-                        tracing::warn!(
-                            tool = name,
-                            path = %cap_path.display(),
-                            "Capabilities file missing \"parameters\" field; \
-                             tool will accept any JSON object (permissive fallback)"
-                        );
-                    }
-                    (caps, oauth, desc, params)
-                } else {
                     tracing::warn!(
+                        tool = name,
                         path = %cap_path.display(),
-                        "Capabilities file not found, using default (no permissions)"
+                        "Capabilities file missing \"description\" field; \
+                         tool will use generic fallback description"
                     );
-                    (Capabilities::default(), None, None, None)
                 }
+                (caps, oauth, desc)
             } else {
                 tracing::warn!(
                     tool = name,
-                    "No capabilities file for WASM tool; \
+                    path = %cap_path.display(),
-                     tool will use generic fallback description and accept any JSON object"
+                    "Capabilities file not found, using default (no permissions)"
                 );
-                (Capabilities::default(), None, None, None)
+                (Capabilities::default(), None, None)
-            };
+            }
+        } else {
+            tracing::warn!(
+                tool = name,
+                "No capabilities file for WASM tool; \
+                     tool will use generic fallback description"
+            );
+            (Capabilities::default(), None, None)
+        };
 
         // Register the tool
         self.registry
@@ -200,7 +178,7 @@ impl WasmToolLoader {
                 capabilities,
                 limits: None,
                 description: description.as_deref(),
-                schema,
+                schema: None,
                 secrets_store: self.secrets_store.clone(),
                 oauth_refresh,
             })
@@ -440,6 +418,7 @@ fn resolve_oauth_refresh_config(cap_file: &CapabilitiesFile) -> Option<OAuthRefr
     let oauth = auth.oauth.as_ref()?;
 
     let builtin = crate::cli::oauth_defaults::builtin_credentials(&auth.secret_name);
+    let exchange_proxy_url = crate::cli::oauth_defaults::exchange_proxy_url();
 
     let client_id = oauth
         .client_id
@@ -462,11 +441,21 @@ fn resolve_oauth_refresh_config(cap_file: &CapabilitiesFile) -> Option<OAuthRefr
                 .and_then(|env| std::env::var(env).ok())
         })
         .or_else(|| builtin.as_ref().map(|c| c.client_secret.to_string()));
+    let client_secret = crate::cli::oauth_defaults::hosted_proxy_client_secret(
+        &client_secret,
+        builtin.as_ref(),
+        exchange_proxy_url.is_some(),
+    );
+    let gateway_token = crate::config::helpers::env_or_override("GATEWAY_AUTH_TOKEN")
+        .map(|token| token.trim().to_string())
+        .filter(|token| !token.is_empty());
 
     Some(OAuthRefreshConfig {
         token_url: oauth.token_url.clone(),
         client_id,
         client_secret,
+        exchange_proxy_url,
+        gateway_token,
         secret_name: auth.secret_name.clone(),
         provider: auth.provider.clone(),
     })
@@ -733,9 +722,44 @@ mod tests {
 
     use tempfile::TempDir;
 
+    use crate::config::helpers::lock_env;
     use crate::testing::credentials::{TEST_OAUTH_CLIENT_ID, TEST_OAUTH_CLIENT_SECRET};
     use crate::tools::wasm::loader::{WasmLoadError, check_wit_version_compat, discover_tools};
 
+    /// Restores a test-scoped env var override on drop.
+    struct EnvVarGuard {
+        key: String,
+        previous: Option<String>,
+    }
+
+    impl Drop for EnvVarGuard {
+        fn drop(&mut self) {
+            // SAFETY: Tests use lock_env() to serialize environment access.
+            unsafe {
+                if let Some(ref value) = self.previous {
+                    std::env::set_var(&self.key, value);
+                } else {
+                    std::env::remove_var(&self.key);
+                }
+            }
+        }
+    }
+
+    fn set_env_var(key: &str, value: Option<&str>) -> EnvVarGuard {
+        let previous = std::env::var(key).ok();
+        // SAFETY: Tests use lock_env() to serialize environment access.
+        unsafe {
+            match value {
+                Some(value) => std::env::set_var(key, value),
+                None => std::env::remove_var(key),
+            }
+        }
+        EnvVarGuard {
+            key: key.to_string(),
+            previous,
+        }
+    }
+
     #[test]
     fn wit_version_compat_none_is_ok() {
         // Pre-versioning extensions (no wit_version declared) should always pass
@@ -893,6 +917,8 @@ mod tests {
             config.client_secret,
             Some(TEST_OAUTH_CLIENT_SECRET.to_string())
         );
+        assert_eq!(config.exchange_proxy_url, None);
+        assert_eq!(config.gateway_token, None);
         assert_eq!(config.secret_name, "google_oauth_token");
         assert_eq!(config.provider, Some("google".to_string()));
     }
@@ -953,6 +979,10 @@ mod tests {
             AuthCapabilitySchema, CapabilitiesFile, OAuthConfigSchema,
         };
 
+        let _guard = lock_env();
+        let _proxy_guard = set_env_var("IRONCLAW_OAUTH_EXCHANGE_URL", None);
+        let _gateway_token_guard = set_env_var("GATEWAY_AUTH_TOKEN", None);
+
         // google_oauth_token should fall back to built-in credentials
         let caps = CapabilitiesFile {
             auth: Some(AuthCapabilitySchema {
@@ -974,6 +1004,95 @@ mod tests {
         let config = config.unwrap();
         assert!(!config.client_id.is_empty());
         assert!(config.client_secret.is_some());
+        assert_eq!(config.exchange_proxy_url, None);
+        assert_eq!(config.gateway_token, None);
+    }
+
+    #[test]
+    fn test_resolve_oauth_refresh_config_hosted_proxy_populates_env_and_suppresses_builtin_secret()
+    {
+        use crate::tools::wasm::capabilities_schema::{
+            AuthCapabilitySchema, CapabilitiesFile, OAuthConfigSchema,
+        };
+
+        let _guard = lock_env();
+        let _proxy_guard = set_env_var(
+            "IRONCLAW_OAUTH_EXCHANGE_URL",
+            Some("https://compose-api.example.com"),
+        );
+        let _gateway_token_guard = set_env_var("GATEWAY_AUTH_TOKEN", Some("gateway-test-token"));
+        let _client_id_guard =
+            set_env_var("GOOGLE_OAUTH_CLIENT_ID", Some("hosted-google-client-id"));
+
+        let caps = CapabilitiesFile {
+            auth: Some(AuthCapabilitySchema {
+                secret_name: "google_oauth_token".to_string(),
+                provider: Some("google".to_string()),
+                oauth: Some(OAuthConfigSchema {
+                    authorization_url: "https://accounts.google.com/o/oauth2/v2/auth".to_string(),
+                    token_url: "https://oauth2.googleapis.com/token".to_string(),
+                    client_id_env: Some("GOOGLE_OAUTH_CLIENT_ID".to_string()),
+                    ..Default::default()
+                }),
+                ..Default::default()
+            }),
+            ..Default::default()
+        };
+
+        let config = super::resolve_oauth_refresh_config(&caps).expect("hosted oauth config");
+        assert_eq!(config.client_id, "hosted-google-client-id");
+        assert_eq!(config.client_secret, None);
+        assert_eq!(
+            config.exchange_proxy_url.as_deref(),
+            Some("https://compose-api.example.com")
+        );
+        assert_eq!(config.gateway_token.as_deref(), Some("gateway-test-token"));
+    }
+
+    #[test]
+    fn test_resolve_oauth_refresh_config_hosted_proxy_preserves_explicit_secret() {
+        use crate::tools::wasm::capabilities_schema::{
+            AuthCapabilitySchema, CapabilitiesFile, OAuthConfigSchema,
+        };
+
+        let _guard = lock_env();
+        let _proxy_guard = set_env_var(
+            "IRONCLAW_OAUTH_EXCHANGE_URL",
+            Some("https://compose-api.example.com"),
+        );
+        let _gateway_token_guard = set_env_var("GATEWAY_AUTH_TOKEN", Some("gateway-test-token"));
+        let _client_id_guard =
+            set_env_var("GOOGLE_OAUTH_CLIENT_ID", Some("hosted-google-client-id"));
+        let _client_secret_guard =
+            set_env_var("GOOGLE_OAUTH_CLIENT_SECRET", Some("hosted-server-secret"));
+
+        let caps = CapabilitiesFile {
+            auth: Some(AuthCapabilitySchema {
+                secret_name: "google_oauth_token".to_string(),
+                provider: Some("google".to_string()),
+                oauth: Some(OAuthConfigSchema {
+                    authorization_url: "https://accounts.google.com/o/oauth2/v2/auth".to_string(),
+                    token_url: "https://oauth2.googleapis.com/token".to_string(),
+                    client_id_env: Some("GOOGLE_OAUTH_CLIENT_ID".to_string()),
+                    client_secret_env: Some("GOOGLE_OAUTH_CLIENT_SECRET".to_string()),
+                    ..Default::default()
+                }),
+                ..Default::default()
+            }),
+            ..Default::default()
+        };
+
+        let config = super::resolve_oauth_refresh_config(&caps).expect("hosted oauth config");
+        assert_eq!(config.client_id, "hosted-google-client-id");
+        assert_eq!(
+            config.client_secret.as_deref(),
+            Some("hosted-server-secret")
+        );
+        assert_eq!(
+            config.exchange_proxy_url.as_deref(),
+            Some("https://compose-api.example.com")
+        );
+        assert_eq!(config.gateway_token.as_deref(), Some("gateway-test-token"));
     }
 
     // ---------------------------------------------------------------

src/tools/wasm/wrapper.rs

@@ -19,7 +19,7 @@ use wasmtime_wasi::{ResourceTable, WasiCtx, WasiCtxBuilder, WasiView};
 use crate::context::JobContext;
 use crate::llm::recording::{HttpExchangeRequest, HttpExchangeResponse, HttpInterceptor};
 use crate::safety::LeakDetector;
-use crate::secrets::SecretsStore;
+use crate::secrets::{DecryptedSecret, SecretsStore};
 use crate::tools::tool::{Tool, ToolError, ToolOutput};
 use crate::tools::wasm::capabilities::Capabilities;
 use crate::tools::wasm::credential_injector::{
@@ -44,6 +44,7 @@ wasmtime::component::bindgen!({
 });
 
 // Alias the export interface types for convenience.
+use crate::cli::oauth_defaults;
 use exports::near::agent::tool as wit_tool;
 
 /// Configuration needed to refresh an expired OAuth access token.
@@ -59,6 +60,10 @@ pub struct OAuthRefreshConfig {
     pub client_id: String,
     /// OAuth client_secret (optional, some providers use PKCE without a secret).
     pub client_secret: Option<String>,
+    /// Hosted OAuth proxy base URL (e.g., "http://host.docker.internal:8080").
+    pub exchange_proxy_url: Option<String>,
+    /// Gateway auth token for authenticating with the hosted OAuth proxy.
+    pub gateway_token: Option<String>,
     /// Secret name of the access token (e.g., "google_oauth_token").
     /// The refresh token lives at `{secret_name}_refresh_token`.
     pub secret_name: String,
@@ -656,12 +661,125 @@ impl WasmToolSchemas {
     }
 
     fn new(discovery: serde_json::Value) -> Self {
+        let advertised = Self::compact_schema(&discovery);
         Self {
-            advertised: Self::permissive_schema(),
+            advertised,
             discovery,
         }
     }
 
+    /// Derive a compact advertised schema from the full discovery schema.
+    ///
+    /// Collects properties from top-level `properties` and from
+    /// `oneOf`/`anyOf`/`allOf` variants. Keeps only properties that are in
+    /// the top-level `required` array or carry an `enum`/`const` constraint.
+    /// For properties defined via `const` across multiple variants (e.g.
+    /// `"action": {"const": "get_repo"}` in each `oneOf` branch), the `const`
+    /// values are merged into a single `enum` array.
+    ///
+    /// Variant-level `required` fields (e.g. `owner`, `repo` required within
+    /// each `oneOf` variant but not top-level) are intentionally omitted from
+    /// the compact schema — the LLM can discover them via
+    /// `tool_info(detail: "schema")`.
+    ///
+    /// At most `MAX_COMPACT_PROPERTIES` properties are collected to bound
+    /// allocations from adversarial schemas.
+    fn compact_schema(discovery: &serde_json::Value) -> serde_json::Value {
+        const MAX_COMPACT_PROPERTIES: usize = 100;
+
+        let required: std::collections::HashSet<String> = discovery
+            .get("required")
+            .and_then(|r| r.as_array())
+            .map(|arr| {
+                arr.iter()
+                    .filter_map(|v| v.as_str().map(String::from))
+                    .collect()
+            })
+            .unwrap_or_default();
+
+        // Collect properties from top-level and oneOf/anyOf/allOf variants.
+        // For properties with `const` across variants, merge into an `enum`.
+        let mut all_properties = serde_json::Map::new();
+        // Track const values per property to merge into enum.
+        let mut const_values: std::collections::HashMap<String, Vec<serde_json::Value>> =
+            std::collections::HashMap::new();
+
+        if let Some(props) = discovery.get("properties").and_then(|p| p.as_object()) {
+            for (k, v) in props {
+                if all_properties.len() >= MAX_COMPACT_PROPERTIES {
+                    break;
+                }
+                all_properties.insert(k.clone(), v.clone());
+            }
+        }
+        for key in ["oneOf", "anyOf", "allOf"] {
+            if let Some(variants) = discovery.get(key).and_then(|v| v.as_array()) {
+                for variant in variants {
+                    if let Some(props) = variant.get("properties").and_then(|p| p.as_object()) {
+                        for (k, v) in props {
+                            if all_properties.len() >= MAX_COMPACT_PROPERTIES
+                                && !all_properties.contains_key(k)
+                            {
+                                continue;
+                            }
+                            // Track const values for merging into enum.
+                            if let Some(c) = v.get("const") {
+                                const_values.entry(k.clone()).or_default().push(c.clone());
+                            }
+                            all_properties.entry(k.clone()).or_insert_with(|| v.clone());
+                        }
+                    }
+                }
+            }
+        }
+
+        // Merge collected const values into enum arrays.
+        for (name, values) in &const_values {
+            if values.len() > 1
+                && let Some(prop) = all_properties.get_mut(name)
+            {
+                let mut merged = prop.clone();
+                if let Some(obj) = merged.as_object_mut() {
+                    obj.remove("const");
+                    obj.insert("enum".to_string(), serde_json::Value::Array(values.clone()));
+                }
+                *prop = merged;
+            }
+        }
+
+        if all_properties.is_empty() {
+            return Self::permissive_schema();
+        }
+
+        let kept: serde_json::Map<String, serde_json::Value> = all_properties
+            .into_iter()
+            .filter(|(name, prop)| {
+                required.contains(name) || prop.get("enum").is_some() || prop.get("const").is_some()
+            })
+            .collect();
+
+        if kept.is_empty() {
+            return Self::permissive_schema();
+        }
+
+        let kept_required: Vec<serde_json::Value> = required
+            .iter()
+            .filter(|name| kept.contains_key(name.as_str()))
+            .map(|name| serde_json::Value::String(name.clone()))
+            .collect();
+
+        let mut result = serde_json::json!({
+            "type": "object",
+            "properties": kept,
+            "additionalProperties": true,
+        });
+        if !kept_required.is_empty() {
+            result["required"] = serde_json::Value::Array(kept_required);
+        }
+
+        result
+    }
+
     fn with_override(&self, schema: serde_json::Value) -> Self {
         Self {
             advertised: schema.clone(),
@@ -1097,6 +1215,53 @@ async fn refresh_oauth_token(
     user_id: &str,
     config: &OAuthRefreshConfig,
 ) -> bool {
+    let refresh_name = format!("{}_refresh_token", config.secret_name);
+
+    if let Some(proxy_url) = config.exchange_proxy_url.as_deref() {
+        let Some(gateway_token) = config.gateway_token.as_deref() else {
+            tracing::warn!(
+                "OAuth refresh proxy is configured, but no gateway auth token is available"
+            );
+            return false;
+        };
+
+        // In hosted mode, the configured exchange proxy owns the outbound token
+        // refresh and validation policy for the provider token_url. Direct-mode
+        // HTTPS/private-IP checks remain in place for self-hosted refreshes below.
+        let refresh_secret = match load_oauth_refresh_secret(store, user_id, &refresh_name).await {
+            Some(secret) => secret,
+            None => return false,
+        };
+        let token_response = match oauth_defaults::refresh_token_via_proxy(
+            oauth_defaults::ProxyRefreshTokenRequest {
+                proxy_url,
+                gateway_token,
+                token_url: &config.token_url,
+                client_id: &config.client_id,
+                client_secret: config.client_secret.as_deref(),
+                refresh_token: refresh_secret.expose(),
+                provider: config.provider.as_deref(),
+            },
+        )
+        .await
+        {
+            Ok(response) => response,
+            Err(error) => {
+                tracing::warn!(error = %error, "OAuth token refresh via proxy failed");
+                return false;
+            }
+        };
+
+        return persist_refreshed_oauth_tokens(
+            store,
+            user_id,
+            config,
+            &refresh_name,
+            token_response,
+        )
+        .await;
+    }
+
     // SSRF defense: token_url comes from the tool's capabilities file.
     if !config.token_url.starts_with("https://") {
         tracing::warn!(
@@ -1114,19 +1279,6 @@ async fn refresh_oauth_token(
         return false;
     }
 
-    let refresh_name = format!("{}_refresh_token", config.secret_name);
-    let refresh_secret = match store.get_decrypted(user_id, &refresh_name).await {
-        Ok(s) => s,
-        Err(e) => {
-            tracing::debug!(
-                secret_name = %refresh_name,
-                error = %e,
-                "No refresh token available, skipping token refresh"
-            );
-            return false;
-        }
-    };
-
     let client = match reqwest::Client::builder()
         .timeout(Duration::from_secs(15))
         .redirect(reqwest::redirect::Policy::none())
@@ -1139,6 +1291,10 @@ async fn refresh_oauth_token(
         }
     };
 
+    let refresh_secret = match load_oauth_refresh_secret(store, user_id, &refresh_name).await {
+        Some(secret) => secret,
+        None => return false,
+    };
     let mut params = vec![
         ("grant_type", "refresh_token".to_string()),
         ("refresh_token", refresh_secret.expose().to_string()),
@@ -1174,22 +1330,55 @@ async fn refresh_oauth_token(
             return false;
         }
     };
-
+    let token_response = match token_data.get("access_token").and_then(|v| v.as_str()) {
-    let new_access_token = match token_data.get("access_token").and_then(|v| v.as_str()) {
+        Some(access_token) => oauth_defaults::OAuthTokenResponse {
-        Some(t) => t,
+            access_token: access_token.to_string(),
+            refresh_token: token_data
+                .get("refresh_token")
+                .and_then(|v| v.as_str())
+                .map(str::to_string),
+            expires_in: token_data.get("expires_in").and_then(|v| v.as_u64()),
+        },
         None => {
             tracing::warn!("Token refresh response missing access_token field");
             return false;
         }
     };
 
-    // Store the new access token with expiry
+    persist_refreshed_oauth_tokens(store, user_id, config, &refresh_name, token_response).await
+}
+
+async fn load_oauth_refresh_secret(
+    store: &(dyn SecretsStore + Send + Sync),
+    user_id: &str,
+    refresh_name: &str,
+) -> Option<DecryptedSecret> {
+    match store.get_decrypted(user_id, refresh_name).await {
+        Ok(secret) => Some(secret),
+        Err(error) => {
+            tracing::debug!(
+                secret_name = %refresh_name,
+                error = %error,
+                "No refresh token available, skipping token refresh"
+            );
+            None
+        }
+    }
+}
+
+async fn persist_refreshed_oauth_tokens(
+    store: &(dyn SecretsStore + Send + Sync),
+    user_id: &str,
+    config: &OAuthRefreshConfig,
+    refresh_name: &str,
+    token_response: oauth_defaults::OAuthTokenResponse,
+) -> bool {
     let mut access_params =
-        crate::secrets::CreateSecretParams::new(&config.secret_name, new_access_token);
+        crate::secrets::CreateSecretParams::new(&config.secret_name, &token_response.access_token);
     if let Some(ref provider) = config.provider {
         access_params = access_params.with_provider(provider);
     }
-    if let Some(expires_in) = token_data.get("expires_in").and_then(|v| v.as_u64()) {
+    if let Some(expires_in) = token_response.expires_in {
         let expires_at = chrono::Utc::now() + chrono::Duration::seconds(expires_in as i64);
         access_params = access_params.with_expiry(expires_at);
     }
@@ -1199,10 +1388,8 @@ async fn refresh_oauth_token(
         return false;
     }
 
-    // Store rotated refresh token if the provider sent a new one
+    if let Some(new_refresh) = token_response.refresh_token.as_deref() {
-    if let Some(new_refresh) = token_data.get("refresh_token").and_then(|v| v.as_str()) {
+        let mut refresh_params = crate::secrets::CreateSecretParams::new(refresh_name, new_refresh);
-        let mut refresh_params =
-            crate::secrets::CreateSecretParams::new(&refresh_name, new_refresh);
         if let Some(ref provider) = config.provider {
             refresh_params = refresh_params.with_provider(provider);
         }
@@ -1551,9 +1738,18 @@ fn build_tool_usage_hint(tool_name: &str, schema: &serde_json::Value) -> String
 
 #[cfg(test)]
 mod tests {
+    use std::collections::HashMap;
+    use std::net::SocketAddr;
     use std::sync::{Arc, Mutex};
 
     use async_trait::async_trait;
+    use axum::extract::{Form, State};
+    use axum::http::HeaderMap;
+    use axum::routing::post;
+    use axum::{Json, Router};
+    use serde_json::json;
+    use tokio::net::TcpListener;
+    use tokio::sync::{Mutex as AsyncMutex, oneshot};
     use uuid::Uuid;
 
     use crate::context::JobContext;
@@ -1643,6 +1839,95 @@ mod tests {
         }
     }
 
+    #[derive(Clone, Debug, PartialEq, Eq)]
+    struct RecordedProxyRequest {
+        authorization: Option<String>,
+        form: HashMap<String, String>,
+    }
+
+    struct MockProxyServer {
+        addr: SocketAddr,
+        requests: Arc<AsyncMutex<Vec<RecordedProxyRequest>>>,
+        shutdown_tx: Option<oneshot::Sender<()>>,
+        server_task: Option<tokio::task::JoinHandle<()>>,
+    }
+
+    impl MockProxyServer {
+        async fn start() -> Self {
+            async fn refresh_handler(
+                State(requests): State<Arc<AsyncMutex<Vec<RecordedProxyRequest>>>>,
+                headers: HeaderMap,
+                Form(form): Form<HashMap<String, String>>,
+            ) -> Json<serde_json::Value> {
+                requests.lock().await.push(RecordedProxyRequest {
+                    authorization: headers
+                        .get(axum::http::header::AUTHORIZATION)
+                        .and_then(|value| value.to_str().ok())
+                        .map(str::to_string),
+                    form,
+                });
+                Json(json!({
+                    "access_token": "mock-refreshed-access-token",
+                    "refresh_token": "mock-rotated-refresh-token",
+                    "expires_in": 3600
+                }))
+            }
+
+            let requests = Arc::new(AsyncMutex::new(Vec::new()));
+            let app = Router::new()
+                .route("/oauth/refresh", post(refresh_handler))
+                .with_state(Arc::clone(&requests));
+
+            let listener = TcpListener::bind("127.0.0.1:0")
+                .await
+                .expect("bind mock proxy");
+            let addr = listener.local_addr().expect("read mock proxy addr");
+            let (shutdown_tx, shutdown_rx) = oneshot::channel::<()>();
+            let server_task = tokio::spawn(async move {
+                let _ = axum::serve(listener, app)
+                    .with_graceful_shutdown(async {
+                        let _ = shutdown_rx.await;
+                    })
+                    .await;
+            });
+
+            Self {
+                addr,
+                requests,
+                shutdown_tx: Some(shutdown_tx),
+                server_task: Some(server_task),
+            }
+        }
+
+        fn base_url(&self) -> String {
+            format!("http://{}", self.addr)
+        }
+
+        async fn requests(&self) -> Vec<RecordedProxyRequest> {
+            self.requests.lock().await.clone()
+        }
+
+        async fn shutdown(mut self) {
+            if let Some(tx) = self.shutdown_tx.take() {
+                let _ = tx.send(());
+            }
+            if let Some(task) = self.server_task.take() {
+                let _ = task.await;
+            }
+        }
+    }
+
+    impl Drop for MockProxyServer {
+        fn drop(&mut self) {
+            if let Some(tx) = self.shutdown_tx.take() {
+                let _ = tx.send(());
+            }
+            if let Some(task) = self.server_task.take() {
+                task.abort();
+            }
+        }
+    }
+
     #[test]
     fn test_wrapper_creation() {
         // This test verifies the runtime can be created
@@ -1655,7 +1940,7 @@ mod tests {
     }
 
     #[tokio::test]
-    async fn test_advertised_schema_stays_permissive_until_sidecar_override() {
+    async fn test_advertised_schema_auto_compacted_from_discovery() {
         let discovery_schema = serde_json::json!({
             "type": "object",
             "properties": {
@@ -1675,42 +1960,7 @@ mod tests {
         wrapper.schemas = super::WasmToolSchemas::new(discovery_schema.clone());
         wrapper.description = "Search documents".to_string();
 
-        // Advertised schema stays permissive; discovery holds the typed schema
+        // Advertised schema is auto-compacted: keeps required props, drops optional
-        assert_eq!(
-            wrapper.parameters_schema(),
-            serde_json::json!({
-                "type": "object",
-                "properties": {},
-                "additionalProperties": true
-            })
-        );
-        assert_eq!(wrapper.discovery_schema(), discovery_schema);
-
-        // Raw description is clean — no tool_info hint baked in
-        assert!(!wrapper.description().contains("tool_info"));
-
-        // But schema() composes the hint at display time when advertised is permissive
-        let schema = wrapper.schema();
-        assert!(
-            schema.description.contains("tool_info"),
-            "schema().description should contain tool_info hint: {}",
-            schema.description
-        );
-        assert!(
-            schema.description.contains("include_schema: true"),
-            "hint should mention include_schema: true: {}",
-            schema.description
-        );
-
-        // After sidecar override, both schemas match and hint disappears
-        let wrapper = wrapper.with_schema(serde_json::json!({
-            "type": "object",
-            "properties": {
-                "query": { "type": "string" }
-            },
-            "required": ["query"]
-        }));
-
         assert_eq!(
             wrapper.parameters_schema(),
             serde_json::json!({
@@ -1718,20 +1968,143 @@ mod tests {
                 "properties": {
                     "query": { "type": "string" }
                 },
-                "required": ["query"]
+                "required": ["query"],
+                "additionalProperties": true
             })
         );
-        assert_eq!(wrapper.discovery_schema(), wrapper.parameters_schema());
+        // Discovery retains the full schema
+        assert_eq!(wrapper.discovery_schema(), discovery_schema);
 
-        // With typed schema, schema() should NOT include tool_info hint
+        // Compacted schema has typed properties, so no tool_info hint needed
         let schema = wrapper.schema();
         assert!(
             !schema.description.contains("tool_info"),
-            "schema().description should not contain tool_info hint when typed: {}",
+            "schema().description should not contain tool_info hint when auto-compacted: {}",
             schema.description
         );
     }
 
+    #[test]
+    fn test_compact_schema_keeps_required_and_enum_properties() {
+        let schema = serde_json::json!({
+            "type": "object",
+            "properties": {
+                "action": {
+                    "type": "string",
+                    "enum": ["list", "get", "create"],
+                    "description": "The operation"
+                },
+                "query": { "type": "string" },
+                "limit": { "type": "integer" },
+                "format": {
+                    "type": "string",
+                    "enum": ["json", "csv"]
+                }
+            },
+            "required": ["action"]
+        });
+
+        let compacted = super::WasmToolSchemas::compact_schema(&schema);
+        let props = compacted["properties"].as_object().unwrap();
+
+        // action: required + enum → kept
+        assert!(props.contains_key("action"));
+        // format: has enum → kept
+        assert!(props.contains_key("format"));
+        // query: not required, no enum → dropped
+        assert!(!props.contains_key("query"));
+        // limit: not required, no enum → dropped
+        assert!(!props.contains_key("limit"));
+        // additionalProperties lets the LLM still pass dropped props
+        assert_eq!(compacted["additionalProperties"], true);
+        assert_eq!(compacted["required"], serde_json::json!(["action"]));
+    }
+
+    #[test]
+    fn test_compact_schema_falls_back_to_permissive_when_empty() {
+        // No required, no enum → permissive fallback
+        let schema = serde_json::json!({
+            "type": "object",
+            "properties": {
+                "query": { "type": "string" },
+                "limit": { "type": "integer" }
+            }
+        });
+
+        let compacted = super::WasmToolSchemas::compact_schema(&schema);
+        assert!(compacted["properties"].as_object().unwrap().is_empty());
+    }
+
+    #[test]
+    fn test_compact_schema_handles_no_properties() {
+        let schema = serde_json::json!({ "type": "object" });
+        let compacted = super::WasmToolSchemas::compact_schema(&schema);
+        assert!(compacted["properties"].as_object().unwrap().is_empty());
+    }
+
+    #[test]
+    fn test_compact_schema_handles_oneof_variants() {
+        // GitHub-style schema: oneOf with no top-level properties, const per variant
+        let schema = serde_json::json!({
+            "type": "object",
+            "required": ["action"],
+            "oneOf": [
+                {
+                    "properties": {
+                        "action": { "const": "get_repo" },
+                        "owner": { "type": "string" },
+                        "repo": { "type": "string" }
+                    },
+                    "required": ["action", "owner", "repo"]
+                },
+                {
+                    "properties": {
+                        "action": { "const": "list_issues" },
+                        "owner": { "type": "string" },
+                        "repo": { "type": "string" },
+                        "state": { "type": "string", "enum": ["open", "closed", "all"] }
+                    },
+                    "required": ["action", "owner", "repo"]
+                }
+            ]
+        });
+
+        let compacted = super::WasmToolSchemas::compact_schema(&schema);
+        let props = compacted["properties"].as_object().unwrap();
+
+        // action: required + const values merged into enum → kept
+        let action = &props["action"];
+        assert!(
+            action.get("enum").is_some(),
+            "action const values should be merged into enum: {action}"
+        );
+        let action_enum = action["enum"].as_array().unwrap();
+        assert!(
+            action_enum.contains(&serde_json::json!("get_repo")),
+            "enum should contain get_repo"
+        );
+        assert!(
+            action_enum.contains(&serde_json::json!("list_issues")),
+            "enum should contain list_issues"
+        );
+        assert!(
+            action.get("const").is_none(),
+            "const should be removed after merging into enum"
+        );
+
+        // state: has enum → kept
+        assert!(
+            props.contains_key("state"),
+            "state should be kept (has enum)"
+        );
+        // owner/repo: not in top-level required, no enum → intentionally dropped
+        // (variant-level required is omitted; discoverable via tool_info)
+        assert!(!props.contains_key("owner"), "owner should be dropped");
+        assert!(!props.contains_key("repo"), "repo should be dropped");
+        assert_eq!(compacted["additionalProperties"], true);
+        assert_eq!(compacted["required"], serde_json::json!(["action"]));
+    }
+
     #[test]
     fn test_capabilities_default() {
         let caps = Capabilities::default();
@@ -1893,8 +2266,6 @@ mod tests {
 
     #[tokio::test]
     async fn test_resolve_host_credentials_bearer() {
-        use std::collections::HashMap;
-
         use crate::secrets::{
             CreateSecretParams, CredentialLocation, CredentialMapping, SecretsStore,
         };
@@ -1940,8 +2311,6 @@ mod tests {
 
     #[tokio::test]
     async fn test_resolve_host_credentials_owner_scope_bearer() {
-        use std::collections::HashMap;
-
         use crate::secrets::{
             CreateSecretParams, CredentialLocation, CredentialMapping, SecretsStore,
         };
@@ -1987,8 +2356,6 @@ mod tests {
 
     #[tokio::test]
     async fn test_execute_resolves_host_credentials_from_owner_scope_context() {
-        use std::collections::HashMap;
-
         use crate::secrets::{CredentialLocation, CredentialMapping};
         use crate::tools::wasm::capabilities::HttpCapability;
 
@@ -2038,8 +2405,6 @@ mod tests {
 
     #[tokio::test]
     async fn test_resolve_host_credentials_missing_secret() {
-        use std::collections::HashMap;
-
         use crate::secrets::{CredentialLocation, CredentialMapping};
         use crate::tools::wasm::capabilities::HttpCapability;
         use crate::tools::wasm::wrapper::resolve_host_credentials;
@@ -2071,8 +2436,6 @@ mod tests {
 
     #[tokio::test]
     async fn test_resolve_host_credentials_skips_refresh_when_not_expired() {
-        use std::collections::HashMap;
-
         use crate::secrets::{
             CreateSecretParams, CredentialLocation, CredentialMapping, SecretsStore,
         };
@@ -2114,6 +2477,8 @@ mod tests {
             token_url: "https://oauth2.googleapis.com/token".to_string(),
             client_id: TEST_OAUTH_CLIENT_ID.to_string(),
             client_secret: Some(TEST_OAUTH_CLIENT_SECRET.to_string()),
+            exchange_proxy_url: None,
+            gateway_token: None,
             secret_name: "google_oauth_token".to_string(),
             provider: Some("google".to_string()),
         };
@@ -2130,8 +2495,6 @@ mod tests {
 
     #[tokio::test]
     async fn test_resolve_host_credentials_skips_refresh_no_config() {
-        use std::collections::HashMap;
-
         use crate::secrets::{
             CreateSecretParams, CredentialLocation, CredentialMapping, SecretsStore,
         };
@@ -2175,8 +2538,6 @@ mod tests {
 
     #[tokio::test]
     async fn test_resolve_host_credentials_skips_refresh_no_expires_at() {
-        use std::collections::HashMap;
-
         use crate::secrets::{
             CreateSecretParams, CredentialLocation, CredentialMapping, SecretsStore,
         };
@@ -2216,6 +2577,8 @@ mod tests {
             token_url: "https://oauth2.googleapis.com/token".to_string(),
             client_id: TEST_OAUTH_CLIENT_ID.to_string(),
             client_secret: Some(TEST_OAUTH_CLIENT_SECRET.to_string()),
+            exchange_proxy_url: None,
+            gateway_token: None,
             secret_name: "google_oauth_token".to_string(),
             provider: Some("google".to_string()),
         };
@@ -2230,6 +2593,249 @@ mod tests {
         );
     }
 
+    #[tokio::test]
+    async fn test_resolve_host_credentials_refreshes_via_proxy_without_direct_token_url_validation()
+    {
+        use crate::secrets::{
+            CreateSecretParams, CredentialLocation, CredentialMapping, SecretsStore,
+        };
+        use crate::tools::wasm::capabilities::HttpCapability;
+        use crate::tools::wasm::wrapper::{OAuthRefreshConfig, resolve_host_credentials};
+
+        let proxy = MockProxyServer::start().await;
+        let store = test_secrets_store();
+
+        store
+            .create(
+                "user1",
+                CreateSecretParams::new("google_oauth_token", "expired-access-token")
+                    .with_expiry(chrono::Utc::now() - chrono::Duration::hours(1)),
+            )
+            .await
+            .unwrap();
+        store
+            .create(
+                "user1",
+                CreateSecretParams::new("google_oauth_token_refresh_token", "stored-refresh-token"),
+            )
+            .await
+            .unwrap();
+
+        let mut credentials = HashMap::new();
+        credentials.insert(
+            "google_oauth_token".to_string(),
+            CredentialMapping {
+                secret_name: "google_oauth_token".to_string(),
+                location: CredentialLocation::AuthorizationBearer,
+                host_patterns: vec!["www.googleapis.com".to_string()],
+            },
+        );
+
+        let caps = Capabilities {
+            http: Some(HttpCapability {
+                credentials,
+                ..Default::default()
+            }),
+            ..Default::default()
+        };
+
+        let oauth_config = OAuthRefreshConfig {
+            token_url: "http://127.0.0.1:9/provider-token-endpoint".to_string(),
+            client_id: "hosted-google-client-id".to_string(),
+            client_secret: None,
+            exchange_proxy_url: Some(proxy.base_url()),
+            gateway_token: Some("gateway-test-token".to_string()),
+            secret_name: "google_oauth_token".to_string(),
+            provider: Some("google".to_string()),
+        };
+
+        let resolved =
+            resolve_host_credentials(&caps, Some(&store), "user1", Some(&oauth_config)).await;
+        assert_eq!(resolved.len(), 1);
+        assert_eq!(
+            resolved[0].headers.get("Authorization"),
+            Some(&"Bearer mock-refreshed-access-token".to_string())
+        );
+
+        let access_secret = store.get("user1", "google_oauth_token").await.unwrap();
+        assert!(
+            access_secret
+                .expires_at
+                .expect("refreshed access token expiry")
+                > chrono::Utc::now()
+        );
+        let access_value = store
+            .get_decrypted("user1", "google_oauth_token")
+            .await
+            .unwrap();
+        assert_eq!(access_value.expose(), "mock-refreshed-access-token");
+
+        let refresh_value = store
+            .get_decrypted("user1", "google_oauth_token_refresh_token")
+            .await
+            .unwrap();
+        assert_eq!(refresh_value.expose(), "mock-rotated-refresh-token");
+
+        let requests = proxy.requests().await;
+        assert_eq!(requests.len(), 1);
+        assert_eq!(
+            requests[0].authorization.as_deref(),
+            Some("Bearer gateway-test-token")
+        );
+        assert_eq!(
+            requests[0].form.get("client_id").map(String::as_str),
+            Some("hosted-google-client-id")
+        );
+        assert_eq!(
+            requests[0].form.get("token_url").map(String::as_str),
+            Some("http://127.0.0.1:9/provider-token-endpoint")
+        );
+        assert_eq!(
+            requests[0].form.get("refresh_token").map(String::as_str),
+            Some("stored-refresh-token")
+        );
+        assert_eq!(
+            requests[0].form.get("provider").map(String::as_str),
+            Some("google")
+        );
+        assert!(!requests[0].form.contains_key("client_secret"));
+
+        proxy.shutdown().await;
+    }
+
+    #[tokio::test]
+    async fn test_resolve_host_credentials_skips_refresh_token_lookup_without_gateway_token() {
+        use crate::secrets::{
+            CreateSecretParams, CredentialLocation, CredentialMapping, SecretsStore,
+        };
+        use crate::tools::wasm::capabilities::HttpCapability;
+        use crate::tools::wasm::wrapper::{OAuthRefreshConfig, resolve_host_credentials};
+
+        let store = RecordingSecretsStore::new();
+
+        store
+            .create(
+                "user1",
+                CreateSecretParams::new("google_oauth_token", "expired-access-token")
+                    .with_expiry(chrono::Utc::now() - chrono::Duration::hours(1)),
+            )
+            .await
+            .unwrap();
+        store
+            .create(
+                "user1",
+                CreateSecretParams::new("google_oauth_token_refresh_token", "stored-refresh-token"),
+            )
+            .await
+            .unwrap();
+
+        let mut credentials = HashMap::new();
+        credentials.insert(
+            "google_oauth_token".to_string(),
+            CredentialMapping {
+                secret_name: "google_oauth_token".to_string(),
+                location: CredentialLocation::AuthorizationBearer,
+                host_patterns: vec!["www.googleapis.com".to_string()],
+            },
+        );
+
+        let caps = Capabilities {
+            http: Some(HttpCapability {
+                credentials,
+                ..Default::default()
+            }),
+            ..Default::default()
+        };
+
+        let oauth_config = OAuthRefreshConfig {
+            token_url: "https://oauth2.googleapis.com/token".to_string(),
+            client_id: "hosted-google-client-id".to_string(),
+            client_secret: None,
+            exchange_proxy_url: Some("https://compose-api.example.com".to_string()),
+            gateway_token: None,
+            secret_name: "google_oauth_token".to_string(),
+            provider: Some("google".to_string()),
+        };
+
+        let resolved =
+            resolve_host_credentials(&caps, Some(&store), "user1", Some(&oauth_config)).await;
+        assert!(resolved.is_empty());
+
+        let lookups = store.decrypted_lookups();
+        assert!(lookups.contains(&("user1".to_string(), "google_oauth_token".to_string())));
+        assert!(!lookups.contains(&(
+            "user1".to_string(),
+            "google_oauth_token_refresh_token".to_string(),
+        )));
+    }
+
+    #[tokio::test]
+    async fn test_resolve_host_credentials_skips_refresh_token_lookup_for_invalid_direct_token_url()
+    {
+        use crate::secrets::{
+            CreateSecretParams, CredentialLocation, CredentialMapping, SecretsStore,
+        };
+        use crate::tools::wasm::capabilities::HttpCapability;
+        use crate::tools::wasm::wrapper::{OAuthRefreshConfig, resolve_host_credentials};
+
+        let store = RecordingSecretsStore::new();
+
+        store
+            .create(
+                "user1",
+                CreateSecretParams::new("google_oauth_token", "expired-access-token")
+                    .with_expiry(chrono::Utc::now() - chrono::Duration::hours(1)),
+            )
+            .await
+            .unwrap();
+        store
+            .create(
+                "user1",
+                CreateSecretParams::new("google_oauth_token_refresh_token", "stored-refresh-token"),
+            )
+            .await
+            .unwrap();
+
+        let mut credentials = HashMap::new();
+        credentials.insert(
+            "google_oauth_token".to_string(),
+            CredentialMapping {
+                secret_name: "google_oauth_token".to_string(),
+                location: CredentialLocation::AuthorizationBearer,
+                host_patterns: vec!["www.googleapis.com".to_string()],
+            },
+        );
+
+        let caps = Capabilities {
+            http: Some(HttpCapability {
+                credentials,
+                ..Default::default()
+            }),
+            ..Default::default()
+        };
+
+        let oauth_config = OAuthRefreshConfig {
+            token_url: "http://127.0.0.1:9/provider-token-endpoint".to_string(),
+            client_id: TEST_OAUTH_CLIENT_ID.to_string(),
+            client_secret: Some(TEST_OAUTH_CLIENT_SECRET.to_string()),
+            exchange_proxy_url: None,
+            gateway_token: None,
+            secret_name: "google_oauth_token".to_string(),
+            provider: Some("google".to_string()),
+        };
+
+        let resolved =
+            resolve_host_credentials(&caps, Some(&store), "user1", Some(&oauth_config)).await;
+        assert!(resolved.is_empty());
+
+        let lookups = store.decrypted_lookups();
+        assert!(lookups.contains(&("user1".to_string(), "google_oauth_token".to_string())));
+        assert!(!lookups.contains(&(
+            "user1".to_string(),
+            "google_oauth_token_refresh_token".to_string(),
+        )));
+    }
+
     #[test]
     fn test_is_private_ip_v4() {
         use std::net::IpAddr;

src/tunnel/cloudflare.rs

@@ -111,10 +111,23 @@ impl Tunnel for CloudflareTunnel {
             }
         }
 
-        // Drain stderr in the background to prevent SIGPIPE/buffer stalls.
+        if let Ok(mut guard) = self.url.write() {
-        tokio::spawn(async move { while let Ok(Some(_)) = reader.next_line().await {} });
+            *guard = Some(public_url.clone());
+        }
 
-        // Drain stdout silently.
+        // We took ownership of cloudflared's stderr pipe above to parse the URL.
+        // cloudflared continues writing logs for its entire lifetime. If we drop
+        // the reader, the pipe closes and cloudflared gets SIGPIPE on its next
+        // write. We can't just store the reader without reading — the OS pipe
+        // buffer fills up and cloudflared blocks. So we drain it in a background
+        // task. The task exits naturally when cloudflared is killed (EOF).
+        let drain_handle = tokio::spawn(async move {
+            while let Ok(Some(line)) = reader.next_line().await {
+                tracing::trace!("cloudflared: {line}");
+            }
+        });
+
+        // Drain stdout silently to prevent SIGPIPE/buffer stalls.
         if let Some(stdout) = stdout {
             tokio::spawn(async move {
                 let mut out_reader = tokio::io::BufReader::new(stdout).lines();
@@ -122,12 +135,11 @@ impl Tunnel for CloudflareTunnel {
             });
         }
 
-        if let Ok(mut guard) = self.url.write() {
-            *guard = Some(public_url.clone());
-        }
-
         let mut guard = self.proc.lock().await;
-        *guard = Some(TunnelProcess { child });
+        *guard = Some(TunnelProcess {
+            child,
+            _pipe_drain: Some(drain_handle),
+        });
 
         Ok(public_url)
     }

src/tunnel/custom.rs

@@ -73,6 +73,7 @@ impl Tunnel for CustomTunnel {
         let stderr = child.stderr.take();
 
         let mut public_url = format!("http://{local_host}:{local_port}");
+        let mut drain_handle: Option<tokio::task::JoinHandle<()>> = None;
 
         if self.url_pattern.is_some()
             && let Some(stdout) = stdout
@@ -103,17 +104,26 @@ impl Tunnel for CustomTunnel {
                     Err(_) => {}
                 }
             }
-            // Drain remaining stdout to prevent SIGPIPE/buffer stalls.
+            // We took ownership of the process's stdout pipe above to parse the
-            tokio::spawn(async move { while let Ok(Some(_)) = reader.next_line().await {} });
+            // URL. The process may continue writing to stdout for its lifetime.
+            // If we drop the reader, the pipe closes and the process gets SIGPIPE.
+            // We can't just store the reader without reading — the OS pipe buffer
+            // fills up and the process blocks. So we drain it in a background task.
+            // The task exits naturally when the process is killed (EOF).
+            drain_handle = Some(tokio::spawn(async move {
+                while let Ok(Some(line)) = reader.next_line().await {
+                    tracing::trace!("custom-tunnel: {line}");
+                }
+            }));
         } else if let Some(stdout) = stdout {
-            // No url_pattern: still drain stdout to prevent pipe stalls.
+            // No url_pattern: still drain stdout to prevent SIGPIPE/buffer stalls.
             tokio::spawn(async move {
                 let mut reader = tokio::io::BufReader::new(stdout).lines();
                 while let Ok(Some(_)) = reader.next_line().await {}
             });
         }
 
-        // Drain stderr silently.
+        // Drain stderr to prevent SIGPIPE/buffer stalls.
         if let Some(stderr) = stderr {
             tokio::spawn(async move {
                 let mut reader = tokio::io::BufReader::new(stderr).lines();
@@ -126,7 +136,10 @@ impl Tunnel for CustomTunnel {
         }
 
         let mut guard = self.proc.lock().await;
-        *guard = Some(TunnelProcess { child });
+        *guard = Some(TunnelProcess {
+            child,
+            _pipe_drain: drain_handle,
+        });
 
         Ok(public_url)
     }

src/tunnel/mod.rs

@@ -66,6 +66,10 @@ pub trait Tunnel: Send + Sync {
 /// Wraps a spawned tunnel child process.
 pub(crate) struct TunnelProcess {
     pub child: tokio::process::Child,
+    /// Background task that drains the process's output pipe (stdout or stderr).
+    /// Must stay alive or the process dies (SIGPIPE from closed pipe) or hangs
+    /// (OS pipe buffer fills up, blocking the process's writes).
+    pub _pipe_drain: Option<tokio::task::JoinHandle<()>>,
 }
 
 pub(crate) type SharedProcess = Arc<Mutex<Option<TunnelProcess>>>;
@@ -182,6 +186,22 @@ pub fn create_tunnel(config: &TunnelProviderConfig) -> Result<Option<Box<dyn Tun
 
 // ── Managed tunnel startup ───────────────────────────────────────
 
+/// Determine which local address the tunnel should forward traffic to.
+///
+/// Prefers the webhook server (`HTTP_PORT`) since that's where webhook routes
+/// (Telegram, etc.) are served. Falls back to the gateway port if configured,
+/// otherwise defaults to 0.0.0.0:8080 (the same fallback the webhook server
+/// uses in main.rs when no HTTP config is present).
+fn resolve_tunnel_target(channels: &crate::config::ChannelsConfig) -> (&str, u16) {
+    if let Some(ref http) = channels.http {
+        return (http.host.as_str(), http.port);
+    }
+    if let Some(ref gw) = channels.gateway {
+        return (gw.host.as_str(), gw.port);
+    }
+    ("0.0.0.0", 8080)
+}
+
 /// Start a managed tunnel if configured and no static URL is already set.
 ///
 /// Returns the (potentially mutated) config with `tunnel.public_url` set,
@@ -201,28 +221,17 @@ pub async fn start_managed_tunnel(
         return (config, None);
     };
 
-    let gateway_port = config
+    let (tunnel_host, tunnel_port) = resolve_tunnel_target(&config.channels);
-        .channels
-        .gateway
-        .as_ref()
-        .map(|g| g.port)
-        .unwrap_or(3000);
-    let gateway_host = config
-        .channels
-        .gateway
-        .as_ref()
-        .map(|g| g.host.as_str())
-        .unwrap_or("127.0.0.1");
 
     match create_tunnel(provider_config) {
         Ok(Some(tunnel)) => {
             tracing::debug!(
                 "Starting {} tunnel on {}:{}...",
                 tunnel.name(),
-                gateway_host,
+                tunnel_host,
-                gateway_port
+                tunnel_port
             );
-            match tunnel.start(gateway_host, gateway_port).await {
+            match tunnel.start(tunnel_host, tunnel_port).await {
                 Ok(url) => {
                     tracing::debug!("Tunnel started: {}", url);
                     config.tunnel.public_url = Some(url);
@@ -383,10 +392,111 @@ mod tests {
 
         {
             let mut guard = proc.lock().await;
-            *guard = Some(TunnelProcess { child });
+            *guard = Some(TunnelProcess {
+                child,
+                _pipe_drain: None,
+            });
         }
 
         kill_shared(&proc).await.unwrap();
         assert!(proc.lock().await.is_none());
     }
+
+    // ── Port selection regression tests ──────────────────────────────
+
+    fn base_channels() -> crate::config::ChannelsConfig {
+        crate::config::ChannelsConfig {
+            cli: crate::config::CliConfig { enabled: false },
+            http: None,
+            gateway: None,
+            signal: None,
+            wasm_channels_dir: std::env::temp_dir().join("ironclaw-test-channels"),
+            wasm_channels_enabled: false,
+            wasm_channel_owner_ids: std::collections::HashMap::new(),
+        }
+    }
+
+    fn channels_with_http(host: &str, port: u16) -> crate::config::ChannelsConfig {
+        let mut c = base_channels();
+        c.http = Some(crate::config::HttpConfig {
+            host: host.to_string(),
+            port,
+            webhook_secret: None,
+            user_id: "test".to_string(),
+        });
+        c.gateway = Some(crate::config::GatewayConfig {
+            host: "127.0.0.1".to_string(),
+            port: 3000,
+            auth_token: None,
+            user_id: "test".to_string(),
+            workspace_read_scopes: Vec::new(),
+            memory_layers: Vec::new(),
+            user_tokens: None,
+        });
+        c
+    }
+
+    fn channels_gateway_only(host: &str, port: u16) -> crate::config::ChannelsConfig {
+        let mut c = base_channels();
+        c.gateway = Some(crate::config::GatewayConfig {
+            host: host.to_string(),
+            port,
+            auth_token: None,
+            user_id: "test".to_string(),
+            workspace_read_scopes: Vec::new(),
+            memory_layers: Vec::new(),
+            user_tokens: None,
+        });
+        c
+    }
+
+    fn channels_neither() -> crate::config::ChannelsConfig {
+        base_channels()
+    }
+
+    #[test]
+    fn tunnel_target_prefers_http_port() {
+        let channels = channels_with_http("0.0.0.0", 8080);
+        let (host, port) = resolve_tunnel_target(&channels);
+        assert_eq!(host, "0.0.0.0"); // safety: test-only
+        assert_eq!(port, 8080); // safety: test-only
+    }
+
+    #[test]
+    fn tunnel_target_falls_back_to_gateway() {
+        let channels = channels_gateway_only("10.0.0.1", 4000);
+        let (host, port) = resolve_tunnel_target(&channels);
+        assert_eq!(host, "10.0.0.1"); // safety: test-only
+        assert_eq!(port, 4000); // safety: test-only
+    }
+
+    #[test]
+    fn tunnel_target_defaults_to_webhook_fallback() {
+        let channels = channels_neither();
+        let (host, port) = resolve_tunnel_target(&channels);
+        // Matches the webhook server's hardcoded fallback in main.rs
+        assert_eq!(host, "0.0.0.0"); // safety: test-only
+        assert_eq!(port, 8080); // safety: test-only
+    }
+
+    #[test]
+    fn tunnel_target_http_takes_priority_over_gateway() {
+        let channels = channels_with_http("192.168.1.1", 9090);
+        let (host, port) = resolve_tunnel_target(&channels);
+        // Should use HTTP config, not gateway's 127.0.0.1:3000
+        assert_eq!(host, "192.168.1.1"); // safety: test-only
+        assert_eq!(port, 9090); // safety: test-only
+    }
+
+    #[test]
+    fn tunnel_target_no_http_no_gateway_matches_webhook_fallback() {
+        // When HTTP_PORT is not set and gateway is not configured (e.g. WASM
+        // channels exist but no explicit HTTP config), the webhook server in
+        // main.rs binds to 0.0.0.0:8080 as a hardcoded fallback. The tunnel
+        // must target the same address so webhook traffic reaches the right
+        // server.
+        let channels = channels_neither();
+        let (host, port) = resolve_tunnel_target(&channels);
+        assert_eq!((host, port), ("0.0.0.0", 8080)); // safety: test-only
+    }
 }

src/tunnel/ngrok.rs

@@ -110,12 +110,24 @@ impl Tunnel for NgrokTunnel {
             }
         }
 
-        // Drain stdout silently — ngrok only emits low-level connection events
+        if let Ok(mut guard) = self.url.write() {
-        // to stdout; the pipe must be consumed to prevent SIGPIPE/buffer stalls.
+            *guard = Some(public_url.clone());
-        tokio::spawn(async move { while let Ok(Some(_)) = reader.next_line().await {} });
+        }
 
-        // Drain stderr silently — with --log stdout all meaningful output goes
+        // We took ownership of ngrok's stdout pipe above to parse the URL.
-        // to stdout; stderr only needs to be consumed to prevent pipe stalls.
+        // ngrok continues writing logs to stdout for its entire lifetime.
+        // If we drop the reader, the pipe closes and ngrok gets SIGPIPE on
+        // its next write → process dies. We can't just store the reader
+        // without reading — the OS pipe buffer (~64KB) fills up and ngrok
+        // blocks. So we drain it in a background task. The task exits
+        // naturally when ngrok is killed (EOF on the pipe).
+        let drain_handle = tokio::spawn(async move {
+            while let Ok(Some(line)) = reader.next_line().await {
+                tracing::trace!("ngrok: {line}");
+            }
+        });
+
+        // Drain stderr silently to prevent SIGPIPE/buffer stalls.
         if let Some(stderr) = stderr {
             tokio::spawn(async move {
                 let mut err_reader = tokio::io::BufReader::new(stderr).lines();
@@ -123,12 +135,11 @@ impl Tunnel for NgrokTunnel {
             });
         }
 
-        if let Ok(mut guard) = self.url.write() {
-            *guard = Some(public_url.clone());
-        }
-
         let mut guard = self.proc.lock().await;
-        *guard = Some(TunnelProcess { child });
+        *guard = Some(TunnelProcess {
+            child,
+            _pipe_drain: Some(drain_handle),
+        });
 
         Ok(public_url)
     }

src/util.rs

@@ -1,5 +1,7 @@
 //! Shared utility functions used across the codebase.
 
+use crate::llm::{ChatMessage, Role};
+
 /// Find the largest valid UTF-8 char boundary at or before `pos`.
 ///
 /// Polyfill for `str::floor_char_boundary` (nightly-only). Use when
@@ -16,6 +18,17 @@ pub fn floor_char_boundary(s: &str, pos: usize) -> usize {
     i
 }
 
+/// Ensure the last message in `messages` is a user-role message.
+///
+/// NEAR AI rejects conversations that don't end with a user message;
+/// Claude 4.6 rejects assistant prefill. Call this before any LLM
+/// completion request to satisfy both requirements.
+pub fn ensure_ends_with_user_message(messages: &mut Vec<ChatMessage>) {
+    if !matches!(messages.last(), Some(m) if m.role == Role::User) {
+        messages.push(ChatMessage::user("Continue."));
+    }
+}
+
 /// Check if an LLM response explicitly signals that a job/task is complete.
 ///
 /// Uses phrase-level matching to avoid false positives from bare words like
@@ -72,7 +85,8 @@ pub fn llm_signals_completion(response: &str) -> bool {
 
 #[cfg(test)]
 mod tests {
-    use crate::util::{floor_char_boundary, llm_signals_completion};
+    use crate::llm::ChatMessage;
+    use crate::util::{ensure_ends_with_user_message, floor_char_boundary, llm_signals_completion};
 
     // ── floor_char_boundary ──
 
@@ -103,6 +117,42 @@ mod tests {
         assert_eq!(floor_char_boundary("", 5), 0);
     }
 
+    // ── ensure_ends_with_user_message ──
+
+    #[test]
+    fn ensure_user_message_injects_when_empty() {
+        let mut msgs: Vec<ChatMessage> = vec![];
+        ensure_ends_with_user_message(&mut msgs);
+        assert_eq!(msgs.len(), 1);
+        assert_eq!(msgs[0].role, crate::llm::Role::User);
+    }
+
+    #[test]
+    fn ensure_user_message_injects_after_assistant() {
+        let mut msgs = vec![ChatMessage::user("hi"), ChatMessage::assistant("hello")];
+        ensure_ends_with_user_message(&mut msgs);
+        assert_eq!(msgs.len(), 3);
+        assert_eq!(msgs[2].role, crate::llm::Role::User);
+    }
+
+    #[test]
+    fn ensure_user_message_injects_after_tool_result() {
+        let mut msgs = vec![
+            ChatMessage::user("run tool"),
+            ChatMessage::tool_result("call_1", "my_tool", "result"),
+        ];
+        ensure_ends_with_user_message(&mut msgs);
+        assert_eq!(msgs.len(), 3);
+        assert_eq!(msgs[2].role, crate::llm::Role::User);
+    }
+
+    #[test]
+    fn ensure_user_message_no_op_when_already_user() {
+        let mut msgs = vec![ChatMessage::user("hello")];
+        ensure_ends_with_user_message(&mut msgs);
+        assert_eq!(msgs.len(), 1);
+    }
+
     // ── llm_signals_completion ──
 
     #[test]

src/worker/container.rs

@@ -151,7 +151,7 @@ Job: {}
 Description: {}
 
 You have tools for shell commands, file operations, and code editing.
-Work independently to complete this job. Report when done."#,
+Work independently to complete this job. When finished, your final message MUST include the phrase "The job is complete" to signal termination."#,
             job.title, job.description
         )));
 
@@ -373,6 +373,10 @@ impl LoopDelegate for ContainerDelegate {
         // Poll for follow-up prompts from the user
         self.poll_and_inject_prompt(reason_ctx).await;
 
+        // Claude 4.6 rejects assistant prefill; NEAR AI rejects any non-user-ending
+        // conversation. Ensure the last message is user-role before calling the LLM.
+        crate::util::ensure_ends_with_user_message(&mut reason_ctx.messages);
+
         // Refresh tools (in case WASM tools were built)
         reason_ctx.available_tools = self.tools.tool_definitions().await;
 

src/worker/job.rs

@@ -18,7 +18,7 @@ use crate::agent::agentic_loop::{
 };
 use crate::agent::scheduler::WorkerMessage;
 use crate::agent::task::TaskOutput;
-use crate::channels::web::types::SseEvent;
+use crate::channels::web::types::ToolDecisionDto;
 use crate::context::{ContextManager, JobState};
 use crate::db::Database;
 use crate::error::Error;
@@ -33,6 +33,7 @@ use crate::tools::rate_limiter::RateLimitResult;
 use crate::tools::{
     ApprovalContext, ToolRegistry, autonomous_unavailable_error, prepare_tool_params, redact_params,
 };
+use ironclaw_common::AppEvent;
 
 /// Shared dependencies for worker execution.
 ///
@@ -48,8 +49,8 @@ pub struct WorkerDeps {
     pub hooks: Arc<HookRegistry>,
     pub timeout: Duration,
     pub use_planning: bool,
-    /// SSE broadcast sender for live job event streaming to the web gateway.
+    /// Broadcast sender for live job event streaming to the web gateway.
-    pub sse_tx: Option<tokio::sync::broadcast::Sender<SseEvent>>,
+    pub sse_tx: Option<Arc<crate::channels::web::sse::SseManager>>,
     /// Approval context for tool execution. When `None`, all non-`Never` tools are
     /// blocked (legacy behavior). When `Some`, the context determines which tools
     /// are pre-approved for autonomous execution.
@@ -138,10 +139,10 @@ impl Worker {
         }
 
         // Broadcast SSE for live web UI updates
-        if let Some(ref tx) = self.deps.sse_tx {
+        if let Some(ref sse) = self.deps.sse_tx {
             let job_id_str = job_id.to_string();
             let event = match event_type {
-                "message" => Some(SseEvent::JobMessage {
+                "message" => Some(AppEvent::JobMessage {
                     job_id: job_id_str,
                     role: data
                         .get("role")
@@ -154,7 +155,7 @@ impl Worker {
                         .unwrap_or("")
                         .to_string(),
                 }),
-                "tool_use" => Some(SseEvent::JobToolUse {
+                "tool_use" => Some(AppEvent::JobToolUse {
                     job_id: job_id_str,
                     tool_name: data
                         .get("tool_name")
@@ -166,7 +167,7 @@ impl Worker {
                         .cloned()
                         .unwrap_or(serde_json::Value::Null),
                 }),
-                "tool_result" => Some(SseEvent::JobToolResult {
+                "tool_result" => Some(AppEvent::JobToolResult {
                     job_id: job_id_str,
                     tool_name: data
                         .get("tool_name")
@@ -179,7 +180,7 @@ impl Worker {
                         .unwrap_or("")
                         .to_string(),
                 }),
-                "status" => Some(SseEvent::JobStatus {
+                "status" => Some(AppEvent::JobStatus {
                     job_id: job_id_str,
                     message: data
                         .get("message")
@@ -187,7 +188,7 @@ impl Worker {
                         .unwrap_or("")
                         .to_string(),
                 }),
-                "result" => Some(SseEvent::JobResult {
+                "result" => Some(AppEvent::JobResult {
                     job_id: job_id_str,
                     status: data
                         .get("status")
@@ -200,10 +201,23 @@ impl Worker {
                         .map(|s| s.to_string()),
                     fallback_deliverable: data.get("fallback_deliverable").cloned(),
                 }),
+                "reasoning" => {
+                    let narrative = data
+                        .get("narrative")
+                        .and_then(|v| v.as_str())
+                        .unwrap_or("")
+                        .to_string();
+                    let decisions = ToolDecisionDto::from_json_array(&data["decisions"]);
+                    Some(AppEvent::JobReasoning {
+                        job_id: job_id_str,
+                        narrative,
+                        decisions,
+                    })
+                }
                 _ => None,
             };
             if let Some(event) = event {
-                let _ = tx.send(event);
+                sse.broadcast(event);
             }
         }
     }
@@ -897,6 +911,11 @@ Report when the job is complete or if you encounter issues you cannot resolve."#
                         id: selection.tool_call_id.clone(),
                         name: selection.tool_name.clone(),
                         arguments: selection.parameters.clone(),
+                        reasoning: if action.reasoning.is_empty() {
+                            None
+                        } else {
+                            Some(action.reasoning.clone())
+                        },
                     }],
                 ));
 
@@ -1232,6 +1251,11 @@ impl<'a> LoopDelegate for JobDelegate<'a> {
     ) -> Option<LoopOutcome> {
         // Refresh tool definitions so newly built tools become visible
         reason_ctx.available_tools = self.worker.tools().tool_definitions().await;
+
+        // Claude 4.6 rejects assistant prefill; NEAR AI rejects any non-user-ending
+        // conversation. Ensure the last message is user-role before calling the LLM.
+        crate::util::ensure_ends_with_user_message(&mut reason_ctx.messages);
+
         None
     }
 
@@ -1352,6 +1376,48 @@ impl<'a> LoopDelegate for JobDelegate<'a> {
             );
         }
 
+        // Emit reasoning event if any tool calls carry reasoning.
+        // Sanitize narrative and per-tool rationale through SafetyLayer
+        // (parity with ChatDelegate in dispatcher.rs).
+        let sanitized_narrative = content
+            .as_deref()
+            .filter(|c| !c.trim().is_empty())
+            .map(|c| {
+                self.worker
+                    .deps
+                    .safety
+                    .sanitize_tool_output("job_narrative", c)
+                    .content
+            })
+            .filter(|c| !c.trim().is_empty())
+            .unwrap_or_default();
+        let decisions: Vec<serde_json::Value> = tool_calls
+            .iter()
+            .filter_map(|tc| {
+                tc.reasoning.as_ref().map(|r| {
+                    let sanitized = self
+                        .worker
+                        .deps
+                        .safety
+                        .sanitize_tool_output("tool_rationale", r)
+                        .content;
+                    serde_json::json!({
+                        "tool_name": tc.name,
+                        "rationale": sanitized,
+                    })
+                })
+            })
+            .collect();
+        if !decisions.is_empty() {
+            self.worker.log_event(
+                "reasoning",
+                serde_json::json!({
+                    "narrative": sanitized_narrative,
+                    "decisions": decisions,
+                }),
+            );
+        }
+
         // Add assistant message with tool_calls (OpenAI protocol)
         reason_ctx
             .messages
@@ -1366,7 +1432,7 @@ impl<'a> LoopDelegate for JobDelegate<'a> {
             .map(|tc| ToolSelection {
                 tool_name: tc.name.clone(),
                 parameters: tc.arguments.clone(),
-                reasoning: String::new(),
+                reasoning: tc.reasoning.clone().unwrap_or_default(),
                 alternatives: vec![],
                 tool_call_id: tc.id.clone(),
             })
@@ -1419,6 +1485,11 @@ fn selections_to_tool_calls(selections: &[ToolSelection]) -> Vec<ToolCall> {
             id: s.tool_call_id.clone(),
             name: s.tool_name.clone(),
             arguments: s.parameters.clone(),
+            reasoning: if s.reasoning.is_empty() {
+                None
+            } else {
+                Some(s.reasoning.clone())
+            },
         })
         .collect()
 }

src/workspace/embedding_cache.rs

@@ -3,14 +3,13 @@
 //! Avoids redundant HTTP calls for identical texts by caching embeddings
 //! in memory keyed by `SHA-256(model_name + "\0" + text)`.
 //!
-//! Follows the same cache pattern as `llm::response_cache::CachedProvider`:
+//! Uses `lru::LruCache` for O(1) insertion, lookup, and eviction.
-//! `HashMap` + `last_accessed` tracking + manual LRU eviction.
 
-use std::collections::HashMap;
+use std::num::NonZeroUsize;
 use std::sync::{Arc, Mutex};
-use std::time::Instant;
 
 use async_trait::async_trait;
+use lru::LruCache;
 use sha2::{Digest, Sha256};
 
 use crate::workspace::embeddings::{EmbeddingError, EmbeddingProvider};
@@ -22,8 +21,7 @@ pub struct EmbeddingCacheConfig {
     ///
     /// Approximate raw embedding payload: `max_entries × dimension × 4 bytes`.
     /// At 10,000 entries × 1536 floats ≈ 58 MB (payload only; actual memory
-    /// is higher due to HashMap buckets, `[u8; 32]` hash keys, `Vec`/`Instant`
+    /// is higher due to per-entry overhead in the linked-list LRU).
-    /// per-entry overhead).
     pub max_entries: usize,
 }
 
@@ -35,11 +33,6 @@ impl Default for EmbeddingCacheConfig {
     }
 }
 
-struct CacheEntry {
-    embedding: Vec<f32>,
-    last_accessed: Instant,
-}
-
 /// Embedding provider wrapper that caches results in memory.
 ///
 /// Thread-safe via `std::sync::Mutex`. The lock is **never held**
@@ -47,8 +40,7 @@ struct CacheEntry {
 /// so a synchronous mutex is cheaper than `tokio::sync::Mutex`.
 pub struct CachedEmbeddingProvider {
     inner: Arc<dyn EmbeddingProvider>,
-    cache: Mutex<HashMap<[u8; 32], CacheEntry>>,
+    cache: Mutex<LruCache<[u8; 32], Vec<f32>>>,
-    config: EmbeddingCacheConfig,
 }
 
 impl CachedEmbeddingProvider {
@@ -56,19 +48,18 @@ impl CachedEmbeddingProvider {
     ///
     /// `config.max_entries` is clamped to at least 1.
     pub fn new(inner: Arc<dyn EmbeddingProvider>, config: EmbeddingCacheConfig) -> Self {
-        let config = EmbeddingCacheConfig {
+        let max_entries = config.max_entries.max(1);
-            max_entries: config.max_entries.max(1),
+        if max_entries > 100_000 {
-        };
-        if config.max_entries > 100_000 {
             tracing::warn!(
-                max_entries = config.max_entries,
+                max_entries,
                 "Embedding cache size exceeds 100,000 entries; memory usage may be significant"
             );
         }
+        // safety: max_entries >= 1 due to .max(1) above
+        let cap = NonZeroUsize::new(max_entries).expect("clamped to >= 1"); // safety: always >= 1
         Self {
             inner,
-            cache: Mutex::new(HashMap::with_capacity(config.max_entries.min(1024))),
+            cache: Mutex::new(LruCache::new(cap)),
-            config,
         }
     }
 
@@ -100,49 +91,6 @@ impl CachedEmbeddingProvider {
         hasher.update(text.as_bytes());
         hasher.finalize().into()
     }
-
-    /// Evict the least-recently-used entry if at capacity (single-entry path).
-    // TODO: O(n) scan per eviction. If max_entries grows large, switch to
-    // an ordered data structure (e.g. `IndexMap` with swap_remove, or a
-    // linked-list LRU like the `lru` crate).
-    fn evict_lru(cache: &mut HashMap<[u8; 32], CacheEntry>, max_entries: usize) {
-        while cache.len() >= max_entries {
-            let oldest_key = cache
-                .iter()
-                .min_by_key(|(_, entry)| entry.last_accessed)
-                .map(|(k, _)| *k);
-
-            if let Some(k) = oldest_key {
-                cache.remove(&k);
-            } else {
-                break;
-            }
-        }
-    }
-
-    /// Evict the `k` oldest entries in O(n) average time via partial selection.
-    ///
-    /// Used by `embed_batch` to avoid the O(n×m) cost of calling
-    /// `evict_lru` per insert.
-    fn evict_k_oldest(cache: &mut HashMap<[u8; 32], CacheEntry>, k: usize) {
-        if k == 0 || cache.is_empty() {
-            return;
-        }
-        if k >= cache.len() {
-            cache.clear();
-            return;
-        }
-        // Partial selection: find the k oldest in O(n) average via
-        // select_nth_unstable_by_key, then remove the first k entries.
-        let mut entries: Vec<([u8; 32], Instant)> = cache
-            .iter()
-            .map(|(key, entry)| (*key, entry.last_accessed))
-            .collect();
-        entries.select_nth_unstable_by_key(k - 1, |(_, t)| *t);
-        for (key, _) in entries.into_iter().take(k) {
-            cache.remove(&key);
-        }
-    }
 }
 
 #[async_trait]
@@ -162,39 +110,32 @@ impl EmbeddingProvider for CachedEmbeddingProvider {
     async fn embed(&self, text: &str) -> Result<Vec<f32>, EmbeddingError> {
         let key = self.cache_key(text);
 
-        // Check cache (short critical section)
+        // Check cache (short critical section). LruCache::get promotes the
+        // entry to most-recently-used automatically.
         {
             let mut guard = self.cache.lock().unwrap_or_else(|e| e.into_inner());
-            if let Some(entry) = guard.get_mut(&key) {
+            if let Some(embedding) = guard.get(&key) {
-                entry.last_accessed = Instant::now();
                 tracing::trace!("embedding cache hit");
-                return Ok(entry.embedding.clone());
+                return Ok(embedding.clone());
             }
         }
         // Lock released before HTTP call.
         // NOTE: Thundering herd — multiple concurrent callers with the same
         // uncached key will each call the inner provider. This is acceptable:
-        // embeddings are idempotent and the last writer wins in the HashMap.
+        // embeddings are idempotent and the last writer wins in the LruCache.
 
         let embedding = self.inner.embed(text).await?;
 
-        // Store result. Re-check under lock: another concurrent caller may
+        // Store result under lock. Re-check first: another concurrent caller
-        // have inserted this key while the lock was released for the HTTP call.
+        // may have already cached this key while the lock was released.
         {
             let mut guard = self.cache.lock().unwrap_or_else(|e| e.into_inner());
-            if let Some(entry) = guard.get_mut(&key) {
+            if guard.get(&key).is_some() {
-                // Thundering herd — another caller already cached it.
+                // Thundering herd — another caller beat us. LruCache::get
-                // Just touch timestamp; skip the clone.
+                // already promoted it to most-recently-used; skip the clone.
-                entry.last_accessed = Instant::now();
+                tracing::trace!("embedding cache: concurrent insert, skipping clone");
             } else {
-                Self::evict_lru(&mut guard, self.config.max_entries);
+                guard.push(key, embedding.clone());
-                guard.insert(
-                    key,
-                    CacheEntry {
-                        embedding: embedding.clone(),
-                        last_accessed: Instant::now(),
-                    },
-                );
             }
         }
 
@@ -214,11 +155,9 @@ impl EmbeddingProvider for CachedEmbeddingProvider {
 
         {
             let mut guard = self.cache.lock().unwrap_or_else(|e| e.into_inner());
-            let now = Instant::now();
             for (i, key) in keys.iter().enumerate() {
-                if let Some(entry) = guard.get_mut(key) {
+                if let Some(embedding) = guard.get(key) {
-                    entry.last_accessed = now;
+                    results[i] = Some(embedding.clone());
-                    results[i] = Some(entry.embedding.clone());
                 } else {
                     miss_indices.push(i);
                 }
@@ -228,7 +167,6 @@ impl EmbeddingProvider for CachedEmbeddingProvider {
 
         if miss_indices.is_empty() {
             tracing::trace!(count = texts.len(), "embedding batch: all cache hits");
-            // All slots populated from cache hits
             return results
                 .into_iter()
                 .enumerate()
@@ -260,29 +198,18 @@ impl EmbeddingProvider for CachedEmbeddingProvider {
             "embedding batch: partial cache"
         );
 
-        // Cache FIRST (clone only the cacheable subset), then move originals
+        // Cache only the last `cap` new embeddings — caching more than the
-        // into results. This avoids cloning capacity-skipped embeddings entirely.
+        // cache capacity wastes clone work on entries that are immediately evicted.
         {
             let mut guard = self.cache.lock().unwrap_or_else(|e| e.into_inner());
-            let cacheable = miss_indices.len().min(self.config.max_entries);
+            let cap = guard.cap().get();
-            let skip = miss_indices.len() - cacheable;
+            let skip = miss_indices.len().saturating_sub(cap);
-            let need_to_evict = (guard.len() + cacheable).saturating_sub(self.config.max_entries);
-            if need_to_evict > 0 {
-                Self::evict_k_oldest(&mut guard, need_to_evict);
-            }
-            let now = Instant::now();
             for (&orig_idx, emb) in miss_indices[skip..].iter().zip(&new_embeddings[skip..]) {
-                guard.insert(
+                guard.push(keys[orig_idx], emb.clone());
-                    keys[orig_idx],
-                    CacheEntry {
-                        embedding: emb.clone(),
-                        last_accessed: now,
-                    },
-                );
             }
         }
 
-        // Move originals into results (zero-copy for all, including cached ones).
+        // Move originals into results (zero-copy).
         for (orig_idx, emb) in miss_indices.iter().copied().zip(new_embeddings) {
             results[orig_idx] = Some(emb);
         }

tests/batch_last_run_status_tests.rs

@@ -0,0 +1,191 @@
+//! Tests for batch_get_last_run_status (#1469 N+1 fix).
+//!
+//! Verifies:
+//! 1. Empty input returns empty map
+//! 2. Returns the most recent run status per routine
+//! 3. Routines with no runs are omitted from result
+//! 4. Multiple routines with different statuses are correctly returned
+
+#[cfg(feature = "libsql")]
+mod tests {
+    use std::sync::Arc;
+
+    use chrono::{Duration, Utc};
+    use uuid::Uuid;
+
+    use ironclaw::agent::routine::{
+        Routine, RoutineAction, RoutineGuardrails, RoutineRun, RunStatus, Trigger,
+    };
+    use ironclaw::db::Database;
+
+    async fn create_test_db() -> (Arc<dyn Database>, tempfile::TempDir) {
+        use ironclaw::db::libsql::LibSqlBackend;
+
+        let temp_dir = tempfile::tempdir().expect("tempdir");
+        let db_path = temp_dir.path().join("test.db");
+        let backend = LibSqlBackend::new_local(&db_path)
+            .await
+            .expect("LibSqlBackend");
+        backend.run_migrations().await.expect("migrations");
+        let db: Arc<dyn Database> = Arc::new(backend);
+        (db, temp_dir)
+    }
+
+    fn make_routine(id: Uuid) -> Routine {
+        Routine {
+            id,
+            name: format!("test-routine-{}", id),
+            description: "Test routine".to_string(),
+            user_id: "default".to_string(),
+            enabled: true,
+            trigger: Trigger::Manual,
+            action: RoutineAction::FullJob {
+                title: "Test job".to_string(),
+                description: "Test description".to_string(),
+                max_iterations: 5,
+            },
+            guardrails: RoutineGuardrails {
+                cooldown: std::time::Duration::from_secs(0),
+                max_concurrent: 1,
+                dedup_window: None,
+            },
+            notify: Default::default(),
+            last_run_at: None,
+            next_fire_at: None,
+            run_count: 0,
+            consecutive_failures: 0,
+            state: serde_json::json!({}),
+            created_at: Utc::now(),
+            updated_at: Utc::now(),
+        }
+    }
+
+    fn make_run(
+        routine_id: Uuid,
+        status: RunStatus,
+        started_at: chrono::DateTime<chrono::Utc>,
+    ) -> RoutineRun {
+        RoutineRun {
+            id: Uuid::new_v4(),
+            routine_id,
+            trigger_type: "manual".to_string(),
+            trigger_detail: None,
+            started_at,
+            completed_at: if status == RunStatus::Running {
+                None
+            } else {
+                Some(Utc::now())
+            },
+            status,
+            result_summary: None,
+            tokens_used: None,
+            job_id: None,
+            created_at: Utc::now(),
+        }
+    }
+
+    #[tokio::test]
+    async fn test_batch_get_last_run_status_empty_input() {
+        let (db, _tmp) = create_test_db().await;
+        let result = db
+            .batch_get_last_run_status(&[])
+            .await
+            .expect("batch query");
+        assert!(result.is_empty());
+    }
+
+    #[tokio::test]
+    async fn test_batch_get_last_run_status_returns_latest() {
+        let (db, _tmp) = create_test_db().await;
+
+        let routine_id = Uuid::new_v4();
+        db.create_routine(&make_routine(routine_id))
+            .await
+            .expect("create routine");
+
+        // Create an older run with Ok status
+        let older_run = make_run(routine_id, RunStatus::Ok, Utc::now() - Duration::hours(2));
+        db.create_routine_run(&older_run)
+            .await
+            .expect("create older run");
+        db.complete_routine_run(older_run.id, RunStatus::Ok, None, None)
+            .await
+            .expect("complete older run");
+
+        // Create a newer run with Attention status
+        let newer_run = make_run(
+            routine_id,
+            RunStatus::Attention,
+            Utc::now() - Duration::hours(1),
+        );
+        db.create_routine_run(&newer_run)
+            .await
+            .expect("create newer run");
+        db.complete_routine_run(newer_run.id, RunStatus::Attention, None, None)
+            .await
+            .expect("complete newer run");
+
+        let result = db
+            .batch_get_last_run_status(&[routine_id])
+            .await
+            .expect("batch query");
+        assert_eq!(result.get(&routine_id), Some(&RunStatus::Attention));
+    }
+
+    #[tokio::test]
+    async fn test_batch_get_last_run_status_omits_routines_without_runs() {
+        let (db, _tmp) = create_test_db().await;
+
+        let with_runs = Uuid::new_v4();
+        let without_runs = Uuid::new_v4();
+        db.create_routine(&make_routine(with_runs))
+            .await
+            .expect("create routine");
+        db.create_routine(&make_routine(without_runs))
+            .await
+            .expect("create routine");
+
+        let run = make_run(with_runs, RunStatus::Ok, Utc::now());
+        db.create_routine_run(&run).await.expect("create run");
+        db.complete_routine_run(run.id, RunStatus::Ok, None, None)
+            .await
+            .expect("complete run");
+
+        let result = db
+            .batch_get_last_run_status(&[with_runs, without_runs])
+            .await
+            .expect("batch query");
+        assert_eq!(result.get(&with_runs), Some(&RunStatus::Ok));
+        assert_eq!(result.get(&without_runs), None);
+    }
+
+    #[tokio::test]
+    async fn test_batch_get_last_run_status_multiple_routines() {
+        let (db, _tmp) = create_test_db().await;
+
+        let r1 = Uuid::new_v4();
+        let r2 = Uuid::new_v4();
+        db.create_routine(&make_routine(r1))
+            .await
+            .expect("create r1");
+        db.create_routine(&make_routine(r2))
+            .await
+            .expect("create r2");
+
+        let run1 = make_run(r1, RunStatus::Running, Utc::now());
+        db.create_routine_run(&run1).await.expect("create run1");
+
+        let run2 = make_run(r2, RunStatus::Failed, Utc::now());
+        db.create_routine_run(&run2).await.expect("create run2");
+        db.complete_routine_run(run2.id, RunStatus::Failed, None, None)
+            .await
+            .expect("complete run2");
+
+        let result = db
+            .batch_get_last_run_status(&[r1, r2])
+            .await
+            .expect("batch query");
+        assert_eq!(result.get(&r1), Some(&RunStatus::Running));
+        assert_eq!(result.get(&r2), Some(&RunStatus::Failed));
+    }
+}

tests/e2e/CLAUDE.md

@@ -53,6 +53,7 @@ HEADED=1 pytest scenarios/
 | `test_skills.py` | Skills tab UI visibility, ClawHub search (skipped if registry unreachable), install + remove lifecycle |
 | `test_sse_reconnect.py` | SSE reconnects after programmatic `eventSource.close()` + `connectSSE()`; history is reloaded after reconnect |
 | `test_tool_approval.py` | Approval card appears, buttons disable on approve/deny, parameters toggle via `page.evaluate("showApproval(...)")`; the waiting-approval regression uses a real HTTP tool call |
+| `test_oauth_refresh.py` | Hosted Gmail OAuth regression: complete setup via `/oauth/callback`, expire the stored access token in libSQL, trigger a real `gmail` tool call through `/api/chat/send`, and verify refresh goes through the mock `/oauth/refresh` proxy without forwarding `client_secret` |
 
 ## `helpers.py`
 
@@ -75,6 +76,7 @@ All fixtures are defined in `tests/e2e/conftest.py`. Running `pytest scenarios/`
 | `ironclaw_binary` | Checks `target/debug/ironclaw`; if absent, runs `cargo build --no-default-features --features libsql` (timeout 600s). |
 | `mock_llm_server` | Starts `mock_llm.py --port 0`, reads the assigned port from stdout, waits for `/v1/models` to return 200. Yields the base URL. |
 | `ironclaw_server` | Starts the ironclaw binary with a minimal env (see below), waits for `/api/health` (timeout 60s). Yields the base URL. On teardown sends **SIGINT** (not SIGTERM) so the tokio ctrl_c handler triggers a graceful shutdown and LLVM coverage data is flushed. |
+| `hosted_oauth_refresh_server` | Starts a second ironclaw instance with a dedicated libSQL DB and `GOOGLE_OAUTH_CLIENT_ID=hosted-google-client-id`, while still pointing `IRONCLAW_OAUTH_EXCHANGE_URL` at `mock_llm.py`. Yields a dict with `base_url`, `db_path`, `gateway_user_id`, and `mock_llm_url` for the hosted refresh regression scenario. |
 | `browser` | Launches a single Chromium instance (headless by default; set `HEADED=1` for headed). Shared across all tests. |
 
 ### Function-scoped fixtures
@@ -100,6 +102,8 @@ EMBEDDING_ENABLED=false, SKILLS_ENABLED=true
 ONBOARD_COMPLETED=true   # prevents setup wizard
 ```
 
+The `hosted_oauth_refresh_server` fixture uses the same baseline, but with its own DB/home tempdirs and `GOOGLE_OAUTH_CLIENT_ID=hosted-google-client-id` so hosted OAuth flows exercise proxy credential injection instead of the baked-in desktop Google app.
+
 The binary is also started with `--no-onboard`. Coverage env vars (`CARGO_LLVM_COV*`, `LLVM_*`, `CARGO_ENCODED_RUSTFLAGS`, `CARGO_INCREMENTAL`) are forwarded from the outer environment when present.
 
 ## Mock LLM (`mock_llm.py`)
@@ -113,6 +117,11 @@ python mock_llm.py --port 0
 
 It serves `POST /v1/chat/completions` (streaming + non-streaming) and `GET /v1/models`. Responses are pattern-matched from `CANNED_RESPONSES` against the last user message. Unmatched messages return `"I understand your request."`. The model name reported is always `"mock-model"`.
 
+It also hosts OAuth test endpoints:
+- `POST /oauth/exchange` for hosted auth-code exchange
+- `POST /oauth/refresh` for hosted refresh-token exchange
+- `GET /__mock/oauth/state` and `POST /__mock/oauth/reset` so HTTP E2E scenarios can assert exact proxy payloads and reset counters between setup and refresh assertions
+
 To add a new canned response:
 ```python
 # In mock_llm.py