Compare commits

..
Author SHA1 Message Date
Henry Park 060ce8de25 fix(agent): surface active extension state to the LLM 2026-03-23 11:47:58 -07:00
4 changed files with 356 additions and 95 deletions
+12
View File
@@ -131,6 +131,18 @@ impl Agent {
}
}
if let Some(extension_manager) = self.deps.extension_manager.as_ref() {
match extension_manager.llm_extension_state_summary().await {
Ok(Some(summary)) => {
reasoning = reasoning.with_extension_state_summary(summary);
}
Ok(None) => {}
Err(e) => {
tracing::debug!("Could not load extension state summary: {}", e);
}
}
}
if let Some(prompt) = system_prompt {
reasoning = reasoning.with_system_prompt(prompt);
}
+36 -83
View File
@@ -102,9 +102,6 @@ impl SessionManager {
/// Resolve an external thread ID to an internal thread.
///
/// Returns the session and thread ID. Creates both if they don't exist.
///
/// 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(
&self,
user_id: &str,
@@ -119,57 +116,51 @@ impl SessionManager {
external_thread_id: external_thread_id.map(String::from),
};
// Parse UUID once outside the lock (if applicable)
let ext_uuid = external_thread_id.and_then(|s| Uuid::parse_str(s).ok());
// Single read lock for both the key lookup and UUID adoption check
let adoptable_uuid = {
// Check if we have a mapping
{
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);
}
}
}
// UUID adoption check (still under the same read lock).
// If external_thread_id is a valid UUID not mapped elsewhere,
// it may be a thread created by chat_new_thread_handler or
// hydrated from DB that we can adopt.
if let Some(ext_uuid) = ext_uuid {
let mapped_elsewhere = thread_map.values().any(|&v| v == ext_uuid);
if !mapped_elsewhere {
Some(ext_uuid)
} else {
None
// Check if external_thread_id is itself a known thread UUID that
// exists in the session but was never registered in the thread_map
// (e.g. created by chat_new_thread_handler or hydrated from DB).
// We only adopt it if no thread_map entry maps to this UUID —
// otherwise it belongs to a different channel scope.
if let Some(ext_tid) = external_thread_id
&& let Ok(ext_uuid) = Uuid::parse_str(ext_tid)
{
let thread_map = self.thread_map.read().await;
let mapped_elsewhere = thread_map.values().any(|&v| v == ext_uuid);
drop(thread_map);
if !mapped_elsewhere {
let sess = session.lock().await;
if sess.threads.contains_key(&ext_uuid) {
drop(sess);
let mut thread_map = self.thread_map.write().await;
// Re-check after acquiring write lock to prevent race condition
// where another task mapped this UUID between our read and write.
if !thread_map.values().any(|&v| v == ext_uuid) {
thread_map.insert(key, ext_uuid);
drop(thread_map);
// Ensure undo manager exists
let mut undo_managers = self.undo_managers.write().await;
undo_managers
.entry(ext_uuid)
.or_insert_with(|| Arc::new(Mutex::new(UndoManager::new())));
return (session, ext_uuid);
}
// If it was mapped elsewhere while we were unlocked, fall through
// to create a new thread, preserving channel isolation.
}
} else {
None
}
}; // Single read lock dropped here
// If we found an adoptable UUID, verify it exists in session and acquire write lock
if let Some(ext_uuid) = adoptable_uuid {
let sess = session.lock().await;
if sess.threads.contains_key(&ext_uuid) {
drop(sess);
let mut thread_map = self.thread_map.write().await;
// Re-check after acquiring write lock to prevent race condition
// where another task mapped this UUID between our read and write.
if !thread_map.values().any(|&v| v == ext_uuid) {
thread_map.insert(key, ext_uuid);
drop(thread_map);
// Ensure undo manager exists
let mut undo_managers = self.undo_managers.write().await;
undo_managers
.entry(ext_uuid)
.or_insert_with(|| Arc::new(Mutex::new(UndoManager::new())));
return (session, ext_uuid);
}
// If mapped elsewhere while unlocked, fall through to create new thread
}
}
@@ -918,44 +909,6 @@ 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};
+216
View File
@@ -1498,6 +1498,52 @@ impl ExtensionManager {
Ok(extensions)
}
/// Build a compact, deterministic extension snapshot for LLM prompt context.
pub async fn llm_extension_state_summary(&self) -> Result<Option<String>, ExtensionError> {
let mut extensions = self.list(None, false).await?;
extensions.sort_by(|a, b| {
llm_extension_sort_key(a.kind)
.cmp(&llm_extension_sort_key(b.kind))
.then_with(|| a.name.cmp(&b.name))
});
let mut channels = Vec::new();
let mut tools = Vec::new();
let mut servers = Vec::new();
for extension in extensions {
let owner_bound = matches!(extension.kind, ExtensionKind::WasmChannel)
&& self.has_wasm_channel_owner_binding(&extension.name).await;
if !(extension.active || extension.authenticated || owner_bound) {
continue;
}
let item = llm_extension_summary_item(&extension, owner_bound);
match extension.kind {
ExtensionKind::WasmChannel | ExtensionKind::ChannelRelay => channels.push(item),
ExtensionKind::WasmTool => tools.push(item),
ExtensionKind::McpServer => servers.push(item),
}
}
let mut lines = Vec::new();
if !channels.is_empty() {
lines.push(format!("- Channels: {}", channels.join("; ")));
}
if !tools.is_empty() {
lines.push(format!("- Tools: {}", tools.join("; ")));
}
if !servers.is_empty() {
lines.push(format!("- MCP servers: {}", servers.join("; ")));
}
if lines.is_empty() {
Ok(None)
} else {
Ok(Some(lines.join("\n")))
}
}
/// Remove an installed extension.
pub async fn remove(&self, name: &str) -> Result<String, ExtensionError> {
Self::validate_extension_name(name)?;
@@ -5613,6 +5659,40 @@ fn combine_install_errors(
}
}
fn llm_extension_sort_key(kind: ExtensionKind) -> u8 {
match kind {
ExtensionKind::WasmChannel | ExtensionKind::ChannelRelay => 0,
ExtensionKind::WasmTool => 1,
ExtensionKind::McpServer => 2,
}
}
fn llm_extension_summary_item(extension: &InstalledExtension, owner_bound: bool) -> String {
let mut states = Vec::new();
if extension.authenticated {
states.push("authenticated".to_string());
}
if extension.active {
states.push("active".to_string());
} else if extension.authenticated {
states.push("inactive".to_string());
}
if owner_bound {
states.push("owner-bound".to_string());
}
if !extension.tools.is_empty() {
let mut tool_names = extension.tools.clone();
tool_names.sort();
states.push(format!("tools: {}", tool_names.join(", ")));
}
if states.is_empty() {
extension.name.clone()
} else {
format!("{} ({})", extension.name, states.join(", "))
}
}
#[cfg(test)]
mod tests {
use std::fmt::Debug;
@@ -6571,6 +6651,142 @@ mod tests {
)
}
#[cfg(feature = "libsql")]
#[tokio::test]
async fn test_llm_extension_state_summary_reports_active_owner_bound_telegram()
-> Result<(), String> {
let dir = tempfile::tempdir().map_err(|err| format!("temp dir: {err}"))?;
let channels_dir = dir.path().join("channels");
std::fs::create_dir_all(&channels_dir).map_err(|err| format!("channels dir: {err}"))?;
std::fs::write(channels_dir.join("telegram.wasm"), b"mock")
.map_err(|err| format!("write wasm: {err}"))?;
std::fs::write(
channels_dir.join("telegram.capabilities.json"),
serde_json::to_vec(&serde_json::json!({
"type": "channel",
"name": "telegram",
"setup": {
"required_secrets": [
{
"name": "telegram_bot_token",
"prompt": "Enter your Telegram Bot API token (from @BotFather)",
"optional": false
}
]
},
"capabilities": {
"channel": {
"allowed_paths": ["/webhook/telegram"]
}
},
"config": {
"owner_id": null
}
}))
.map_err(|err| format!("serialize capabilities: {err}"))?,
)
.map_err(|err| format!("write capabilities: {err}"))?;
let (db, _db_tmp) = crate::testing::test_db().await;
let manager = {
use crate::secrets::{InMemorySecretsStore, SecretsCrypto};
use crate::testing::credentials::TEST_CRYPTO_KEY;
use crate::tools::ToolRegistry;
use crate::tools::mcp::process::McpProcessManager;
use crate::tools::mcp::session::McpSessionManager;
let master_key = secrecy::SecretString::from(TEST_CRYPTO_KEY.to_string());
let crypto = Arc::new(
SecretsCrypto::new(master_key)
.map_err(|err| format!("failed to construct test crypto: {err}"))?,
);
ExtensionManager::new(
Arc::new(McpSessionManager::new()),
Arc::new(McpProcessManager::new()),
Arc::new(InMemorySecretsStore::new(crypto)),
Arc::new(ToolRegistry::new()),
None,
None,
dir.path().join("tools"),
channels_dir.clone(),
None,
"test".to_string(),
Some(db),
Vec::new(),
)
};
let channel_manager = Arc::new(ChannelManager::new());
let runtime = Arc::new(
WasmChannelRuntime::new(WasmChannelRuntimeConfig::for_testing())
.map_err(|err| format!("runtime: {err}"))?,
);
let pairing_store = Arc::new(PairingStore::with_base_dir(
dir.path().join("pairing-state"),
));
let router = Arc::new(WasmChannelRouter::new());
manager
.set_channel_runtime(
Arc::clone(&channel_manager),
Arc::clone(&runtime),
Arc::clone(&pairing_store),
Arc::clone(&router),
std::collections::HashMap::new(),
)
.await;
manager
.set_test_wasm_channel_loader(Arc::new({
let runtime = Arc::clone(&runtime);
let pairing_store = Arc::clone(&pairing_store);
move |name| {
Ok(make_test_loaded_channel(
Arc::clone(&runtime),
name,
Arc::clone(&pairing_store),
))
}
}))
.await;
manager
.set_test_telegram_binding_resolver(Arc::new(|_token, existing_owner_id| {
if existing_owner_id.is_some() {
return Err(ExtensionError::Other(
"owner binding should be derived during setup".to_string(),
));
}
Ok(TelegramBindingResult::Bound(TelegramBindingData {
owner_id: 424242,
bot_username: Some("test_hot_bot".to_string()),
binding_state: TelegramOwnerBindingState::VerifiedNow,
}))
}))
.await;
manager
.configure(
"telegram",
&std::collections::HashMap::from([(
"telegram_bot_token".to_string(),
"123456789:ABCdefGhI".to_string(),
)]),
&std::collections::HashMap::new(),
)
.await
.map_err(|err| format!("configure succeeds: {err}"))?;
let summary = manager
.llm_extension_state_summary()
.await
.map_err(|err| format!("summary: {err}"))?
.ok_or_else(|| "expected extension summary".to_string())?;
require(
summary.contains("- Channels: telegram (authenticated, active, owner-bound)"),
format!("unexpected summary: {summary}"),
)
}
#[tokio::test]
async fn test_telegram_hot_activation_returns_verification_challenge_before_binding()
-> Result<(), String> {
+92 -12
View File
@@ -353,6 +353,8 @@ pub struct Reasoning {
workspace_system_prompt: Option<String>,
/// Optional skill context block to inject into system prompt.
skill_context: Option<String>,
/// Optional snapshot of connected/active extensions for the current user.
extension_state_summary: Option<String>,
/// Channel name (e.g. "discord", "telegram") for formatting hints.
channel: Option<String>,
/// Model name for runtime context.
@@ -371,6 +373,7 @@ impl Reasoning {
llm,
workspace_system_prompt: None,
skill_context: None,
extension_state_summary: None,
channel: None,
model_name: None,
is_group_chat: false,
@@ -400,6 +403,14 @@ impl Reasoning {
self
}
/// Set extension runtime context to inject into the system prompt.
pub fn with_extension_state_summary(mut self, summary: String) -> Self {
if !summary.is_empty() {
self.extension_state_summary = Some(summary);
}
self
}
/// Set the channel name for channel-specific formatting hints.
pub fn with_channel(mut self, channel: impl Into<String>) -> Self {
let ch = channel.into();
@@ -932,21 +943,54 @@ Example:
}
fn build_extensions_section_for_tools(&self, tools: &[ToolDefinition]) -> String {
// Only include when the extension management tools are available
let has_ext_tools = tools.iter().any(|t| t.name == "tool_search");
if !has_ext_tools {
let has_search = tools.iter().any(|t| t.name == "tool_search");
let has_list = tools.iter().any(|t| t.name == "tool_list");
let has_info = tools.iter().any(|t| t.name == "extension_info");
if self.extension_state_summary.is_none() && !has_search && !has_list && !has_info {
return String::new();
}
"\n\n## Extensions\n\
You can search, install, and activate extensions to add new capabilities:\n\
- **Channels** (Telegram, Slack, Discord) — messaging integrations. \
When users ask about connecting a messaging platform, search for it as a channel.\n\
- **Tools** — sandboxed functions that extend your abilities.\n\
- **MCP servers** — external API integrations via the Model Context Protocol.\n\n\
Use `tool_search` to find extensions by name. Refer to them by their kind \
(channel, tool, or server) — not as \"MCP server\" generically."
.to_string()
let mut blocks = Vec::new();
if let Some(ref summary) = self.extension_state_summary {
blocks.push(format!(
"Current extension state for this user:\n{}",
summary
));
}
if has_search || has_list || has_info {
let mut guidance = String::from(
"You can search, install, and activate extensions to add new capabilities:\n\
- **Channels** (Telegram, Slack, Discord) — messaging integrations. \
When users ask about connecting a messaging platform, search for it as a channel.\n\
- **Tools** — sandboxed functions that extend your abilities.\n\
- **MCP servers** — external API integrations via the Model Context Protocol.",
);
if has_list {
guidance.push_str(
"\n\nBefore telling the user to connect, activate, or re-enable an extension, \
inspect the current state with `tool_list`.",
);
}
if has_info {
guidance.push_str(
"\nUse `extension_info` when you need deeper compatibility or runtime details \
for an installed extension.",
);
}
if has_search {
guidance.push_str(
"\nUse `tool_search` to find extensions by name. Refer to them by their kind \
(channel, tool, or server) — not as \"MCP server\" generically.",
);
}
blocks.push(guidance);
}
format!("\n\n## Extensions\n{}", blocks.join("\n\n"))
}
fn build_channel_section(&self) -> String {
@@ -2285,6 +2329,42 @@ That's my plan."#;
);
}
#[test]
fn test_system_prompt_includes_extension_runtime_summary() {
let reasoning = make_test_reasoning().with_extension_state_summary(
"- Channels: telegram (authenticated, active, owner-bound)".to_string(),
);
let prompt = reasoning.build_system_prompt_with_tools(&[]);
assert!(
prompt.contains("## Extensions"),
"Prompt should contain an Extensions section when runtime state is present"
);
assert!(
prompt.contains("telegram (authenticated, active, owner-bound)"),
"Prompt should include the injected extension runtime summary"
);
}
#[test]
fn test_system_prompt_extension_guidance_prefers_inspection_before_reconnect() {
let reasoning = make_test_reasoning();
let prompt = reasoning.build_system_prompt_with_tools(&make_tools(&[
"tool_search",
"tool_list",
"extension_info",
]));
assert!(
prompt.contains("inspect the current state with `tool_list`"),
"Prompt should direct the model to inspect installed/active state first"
);
assert!(
prompt.contains("Use `extension_info` when you need deeper compatibility"),
"Prompt should mention extension_info for deeper extension details"
);
}
// ---- plan/evaluate bypass clean_response (Bug #564-2) ----
#[test]