Address tri-agent review findings

stub-patterns.md:
- Add false-positive caveat for return nil/null/{}/[]
- Fix grep portability: use -E flag instead of \s for macOS compat
- Remove fragile Python except pipe pattern
- Add *.tsx/*.jsx to TODO, hardcoded-value, and catch-block patterns
- Add overall "matches require manual review" disclaimer

self-audit.md:
- Run stub detection unconditionally, not gated on TODO count

feature.md:
- Support multiple domains per feature (auth + API + UI)
- Skip probing when no domain section matches

debug-checklists.md:
- Merge duplicate Quick Symptom Lookup into Performance Bugs
  checklist to eliminate DRY violation with Symptom Triage table

Author: 5uck1ess

commands/feature.md

@@ -8,7 +8,7 @@ Complete feature lifecycle: brainstorm → plan → implement → test → lint
 
 ## Step 1: Brainstorm
 
-Identify the feature's domain (auth, API, database, UI, etc.) and read the matching section from `references/domain-probes.md` to surface gray areas. Use the probes to ask targeted questions via `AskUserQuestion` — only for genuinely ambiguous decisions, not obvious ones. Skip probing if the user's input already resolves the gray areas.
+Identify the feature's relevant domains (auth, API, database, UI, etc. — features often span multiple) and read each matching section from `references/domain-probes.md` to surface gray areas. If no section matches, skip domain probing. Use the probes to ask targeted questions via `AskUserQuestion` — only for genuinely ambiguous decisions, not obvious ones. Skip probing if the user's input already resolves the gray areas.
 
 ```
 Feature: {user's input}

commands/references/debug-checklists.md

@@ -46,15 +46,8 @@ When the bug domain isn't obvious, use the symptom to route to the right checkli
 
 ## Performance Bugs
 - Profile first — the slow part is almost never where you think
-- Check for N+1 queries, missing indexes, unbounded loops
-- Memory leak? Compare heap snapshots over time
-- Connection pool or thread pool exhaustion?
-
-### Quick Symptom Lookup
-
-| Symptom | Likely Cause | Investigation |
-|---------|--------------|---------------|
-| Slow API response | N+1 queries | Log SQL count per request |
-| Slow page render | Expensive recomputation | Profile render cycle |
-| Gradual memory growth | Leak (listeners, connections) | Heap snapshots over time |
-| Intermittent slowness | Lock contention / pool exhaustion | Connection pool metrics |
+- Slow API response? Log SQL count per request — likely N+1 queries
+- Slow page render? Profile the render cycle — likely expensive recomputation
+- Gradual memory growth? Heap snapshots over time — likely leaked listeners or connections
+- Intermittent slowness? Check connection pool metrics — likely lock contention or pool exhaustion
+- Check for missing indexes and unbounded loops

commands/references/stub-patterns.md

@@ -2,27 +2,30 @@
 
 Mechanical grep patterns for detecting unfinished work. Use in self-audit's Stale Code section or tri-review to flag incomplete implementations.
 
+All matches require manual review — these are heuristics, not definitive indicators.
+
 ## TODO/FIXME Comments
 
 ```bash
 grep -rn 'TODO\|FIXME\|HACK\|XXX\|PLACEHOLDER' \
-  --include='*.go' --include='*.ts' --include='*.js' \
+  --include='*.go' --include='*.ts' --include='*.tsx' --include='*.js' --include='*.jsx' \
   --include='*.py' --include='*.rs' --include='*.rb' . 2>/dev/null
 ```
 
 ## Empty or Trivial Implementations
 
 ```bash
-# Functions that return nothing useful
-grep -rn 'return nil$\|return null\|return undefined\|return {}\|return \[\]' \
+# Functions that return nothing useful — high false-positive rate.
+# Filter results manually: return nil/null/{}/[] is often legitimate.
+# Strongest signal when combined with TODO/FIXME nearby.
+grep -rn 'return nil\|return null\|return undefined' \
   --include='*.go' --include='*.ts' --include='*.js' --include='*.py' . 2>/dev/null
 
 # Python pass-only functions
-grep -rn '^\s*pass$' --include='*.py' . 2>/dev/null
+grep -rn 'pass$' --include='*.py' . 2>/dev/null
 
-# Empty catch/error blocks
-grep -rn 'catch.*{}\|catch.*{\s*}' --include='*.ts' --include='*.js' . 2>/dev/null
-grep -rn 'except.*:\s*$' --include='*.py' -A1 . 2>/dev/null | grep 'pass'
+# Empty catch/error blocks (use grep -E for extended regex portability)
+grep -rEn 'catch[^{]*\{\s*\}' --include='*.ts' --include='*.tsx' --include='*.js' --include='*.jsx' . 2>/dev/null
 ```
 
 ## Placeholder Text
@@ -35,19 +38,19 @@ grep -rni 'lorem ipsum\|coming soon\|under construction\|placeholder' \
 
 # Template brackets left in
 grep -rn '\[TODO\]\|<TODO>\|{TODO}' \
-  --include='*.md' --include='*.ts' --include='*.go' . 2>/dev/null
+  --include='*.md' --include='*.ts' --include='*.tsx' --include='*.go' . 2>/dev/null
 ```
 
 ## Hardcoded Values Where Dynamic Expected
 
 ```bash
 # Hardcoded IDs or secrets (not in test files)
 grep -rn 'api_key\s*=\s*"[^"]\+"\|password\s*=\s*"[^"]\+"' \
-  --include='*.go' --include='*.ts' --include='*.py' . 2>/dev/null | grep -v '_test\.\|\.test\.\|\.spec\.'
+  --include='*.go' --include='*.ts' --include='*.tsx' --include='*.py' . 2>/dev/null | grep -v '_test\.\|\.test\.\|\.spec\.'
 
 # Hardcoded URLs (not config/const files)
 grep -rn 'http://localhost\|127\.0\.0\.1' \
-  --include='*.go' --include='*.ts' --include='*.py' . 2>/dev/null | grep -v 'config\|const\|test\|spec'
+  --include='*.go' --include='*.ts' --include='*.tsx' --include='*.py' . 2>/dev/null | grep -v 'config\|const\|test\|spec'
 ```
 
 ## Disabled or Skipped Tests

commands/self-audit.md

@@ -91,7 +91,7 @@ grep -rn 'api_key\s*=\s*"[^"]*"' --include='*.go' --include='*.ts' --include='*.
 grep -rn 'TODO\|FIXME\|HACK\|XXX' --include='*.go' --include='*.ts' --include='*.py' --include='*.rs' . 2>/dev/null | wc -l
 ```
 
-For deeper stub detection (empty implementations, placeholder text, hardcoded values, skipped tests), run the patterns from `references/stub-patterns.md`. Only use these when the basic TODO count suggests stale code worth investigating.
+For deeper stub detection (empty implementations, placeholder text, hardcoded values, skipped tests), run the patterns from `references/stub-patterns.md`. Run unconditionally — a repo can have zero TODOs but still contain empty catch blocks, placeholder UI copy, or hardcoded localhost URLs.
 
 ### Documentation