chore: add reviewer-feedback guardrails (CLAUDE.md, pre-commit hook, skill) (#665)

* chore: add reviewer-feedback guardrails (CLAUDE.md, pre-commit hook, skill)

Analysis of ~50 PRs from the past week identified 10 recurring themes
in Copilot and Gemini code review comments. This change addresses them
at development time through three layers:

1. CLAUDE.md additions (7 new rules):
   - Transaction safety for multi-step DB operations
   - UTF-8 string safety (no byte-index slicing)
   - Case-insensitive comparisons for paths/media types
   - Decorator/wrapper trait method delegation
   - Sensitive data redaction in logs/SSE
   - tempfile crate for test temporary files
   - Trust boundaries for worker container data

2. Pre-commit hook (scripts/pre-commit-safety.sh):
   Mechanical checks for unsafe byte slicing, case-sensitive
   extension comparisons, hardcoded /tmp paths, unredacted
   tool parameter logging, and non-transactional DB operations.
   Installed via dev-setup.sh alongside existing commit-msg hook.

3. Review checklist skill (skills/review-checklist/SKILL.md):
   Activates on "review"/"merge" keywords. Covers the judgment-based
   items that can't be linted: transaction safety, SSRF validation,
   approval checks, decorator delegation, test quality, and doc accuracy.

[skip-regression-check]

Co-Authored-By: Claude Opus 4.6 <[email protected]>

* fix: address PR review feedback on pre-commit-safety.sh

- Cache diff output in variable to avoid ~10 redundant git diff calls (Gemini)
- Add early exit when no .rs files are changed (Gemini)
- Fix header comment: list all 5 checks, not just 4 (Copilot)
- Fix check 2 comment: only mentions file extensions, not media types (Copilot)
- Add resolve_base_ref() with fallback candidates instead of hardcoded
  origin/main for standalone mode (Copilot)
- TX check: use -W (function context) to reduce false positives, honor
  // safety: suppression, print triggering lines (Copilot)

[skip-regression-check]

Co-Authored-By: Claude Opus 4.6 <[email protected]>

---------

Co-authored-by: Claude Opus 4.6 <[email protected]>

skills/review-checklist/SKILL.md

@@ -0,0 +1,54 @@
+---
+name: review-checklist
+version: 0.1.0
+description: Pre-merge review checklist based on recurring AI reviewer feedback patterns
+activation:
+  patterns:
+    - "review.*checklist"
+    - "ready to merge"
+    - "pre-merge check"
+    - "check.*before.*merge"
+  keywords:
+    - review
+    - checklist
+    - merge
+    - pre-merge
+  max_context_tokens: 1500
+---
+
+# Pre-Merge Review Checklist
+
+Before merging, verify these items. They represent the most common issues caught by automated code reviewers (Copilot, Gemini) on IronClaw PRs.
+
+## Database Operations
+- [ ] Multi-step DB operations are wrapped in transactions (INSERT+INSERT, UPDATE+DELETE, read-modify-write)
+- [ ] Both postgres AND libsql backends updated for any new Database trait methods
+- [ ] Migrations are atomic (SQL execution + version recording in same transaction)
+
+## Security & Data Safety
+- [ ] Tool parameters are redacted via `redact_params()` before logging or SSE/WebSocket broadcast
+- [ ] URL validation resolves DNS before checking for private/loopback IPs (anti-SSRF via DNS rebinding)
+- [ ] Destructive tools have `requires_approval()` returning `Always` or `UnlessAutoApproved`
+- [ ] Data from worker containers is treated as untrusted (tool domain checks, server-side nesting depth)
+- [ ] No secrets or credentials in error messages, logs, or SSE events
+
+## String Safety
+- [ ] No byte-index slicing (`&s[..n]`) on external/user strings -- use `is_char_boundary()` or `char_indices()`
+- [ ] File extension and media type comparisons are case-insensitive (`.to_ascii_lowercase()` before matching)
+- [ ] Path comparisons are case-insensitive where needed (macOS/Windows filesystems)
+
+## Trait Wrappers & Decorator Chain
+- [ ] New `LlmProvider` trait methods are delegated in ALL wrapper types (grep `impl LlmProvider for`)
+- [ ] New trait methods are tested through the full decorator/provider chain, not just the base impl
+- [ ] Default trait method implementations are intentional -- wrappers that silently return defaults are bugs
+
+## Tests
+- [ ] Temporary files/dirs use `tempfile` crate, no hardcoded `/tmp/` paths
+- [ ] Tests don't mutate global statics without synchronization (use per-test state or `serial_test`)
+- [ ] Tests don't make real network requests (use mocks, stubs, or RFC 5737 TEST-NET IPs like 192.0.2.1)
+- [ ] Test names and comments match actual test behavior and assertions
+
+## Comments & Documentation
+- [ ] Code comments match actual behavior (especially route paths, tool names, function semantics)
+- [ ] Spec/README files updated if module behavior changed
+- [ ] Error messages are clear and non-redundant (don't nest tool name inside tool error that already contains it)