proagents 1.6.9 → 1.6.11

package/.proagents/docs/command-details.md

@@ -0,0 +1,1115 @@
+# ProAgents Command Details
+
+Detailed execution instructions for all `pa:` commands.
+
+---
+
+## CRITICAL: AI Must Execute All Commands
+
+**NEVER tell user to do something. AI must DO it.**
+
+### Commands That READ Data - AI Must Run:
+
+| Command | AI Runs This |
+|---------|--------------|
+| `pa:history` | `grep -v "^#" .proagents/activity.log \| tail -30` |
+| `pa:progress` | `cat .proagents/active-features/_index.json` |
+| `pa:activity` | `cat .proagents/activity.log \| tail -20` |
+| `pa:status` | `cat .proagents/active-features/_index.json` |
+| `pa:context` | `cat .proagents/context.md` |
+| `pa:decisions` | `cat .proagents/decisions.md` |
+| `pa:errors` | `cat .proagents/errors.md` |
+| `pa:feedback` | `cat .proagents/feedback.md` |
+| `pa:handoff-read` | `cat .proagents/handoff.md` |
+| `pa:config` | `cat proagents.config.yaml` |
+| `pa:lock` | `cat .proagents/.lock 2>/dev/null` |
+| `pa:feature-list` | `cat .proagents/active-features/_index.json` |
+| `pa:todo` | `grep -rn "TODO\|FIXME" src/` |
+| `pa:deps` | `cat package.json \| grep dependencies -A 50` |
+| `pa:deps-outdated` | `npm outdated 2>/dev/null` |
+
+### Commands That RUN Tools - AI Must Execute:
+
+| Command | AI Runs This |
+|---------|--------------|
+| `pa:test` | `npm test` (or from config) |
+| `pa:test-unit` | `npm run test:unit` |
+| `pa:test-e2e` | `npm run test:e2e` |
+| `pa:test-mobile` | Install Maestro if missing → Create tests → Run `maestro test` |
+| `pa:lint` | `npm run lint` |
+| `pa:qa` | `npm run lint && npm test` |
+| `pa:coverage` | `npm run test:coverage` |
+| `pa:logs` | `adb logcat -d '*:S' ReactNativeJS:V \| tail -100` |
+| `pa:env-check` | `node -v && npm -v && git --version` |
+| `pa:secrets-scan` | `grep -rn "password\|secret\|api_key" src/` |
+| `pa:db-migrate` | `npm run db:migrate` (or from config) |
+
+### Commands That GENERATE Files - AI Must Create:
+
+**AI MUST create actual files, never just show templates or options.**
+
+| Command | AI Creates | Location |
+|---------|------------|----------|
+| `pa:doc` | Full documentation | `./docs/*.md` |
+| `pa:doc-api` | API documentation | `./docs/api/*.md` |
+| `pa:doc-module X` | Module documentation | `./docs/modules/X.md` |
+| `pa:doc-file X` | File documentation | `./docs/files/X.md` |
+| `pa:changelog` | Updates changelog | `./CHANGELOG.md` |
+| `pa:release` | Release notes | `./RELEASE_NOTES.md` |
+| `pa:readme` | Updates README | `./README.md` |
+| `pa:handoff` | Handoff notes | `.proagents/handoff.md` |
+| `pa:generate-component` | Component file | Based on project structure |
+| `pa:generate-test` | Test file | Based on project structure |
+
+**Example - pa:doc execution:**
+```bash
+# AI creates directory structure
+mkdir -p docs docs/api docs/modules docs/components
+
+# AI analyzes code and creates doc files
+# (AI writes actual content to each file)
+
+# AI reports:
+echo "Created: ./docs/README.md"
+echo "Created: ./docs/api/endpoints.md"
+echo "Total: 8 documentation files"
+```
+
+### Commands That LOG Activity - AI Must Append:
+
+After EVERY pa: command, AI runs:
+```bash
+echo "[$(date '+%Y-%m-%d %H:%M')] [AI_NAME] [COMMAND] Result" >> .proagents/activity.log
+```
+
+---
+
+## Feature Commands
+
+### pa:feature "name"
+
+Start a new feature workflow:
+
+1. Create feature folder: `./.proagents/active-features/feature-[name]/`
+2. Create `status.json`:
+   ```json
+   {
+     "name": "feature-name",
+     "started": "2024-03-06T15:00:00Z",
+     "phase": "analysis",
+     "progress": 0,
+     "branch": "feature/feature-name"
+   }
+   ```
+3. Update `_index.json` to add feature to `active_features`
+4. Create git branch if git enabled
+5. Start analysis phase
+
+### pa:feature-list
+
+**AI reads and displays features:**
+
+```bash
+# AI executes:
+cat .proagents/active-features/_index.json 2>/dev/null
+ls .proagents/active-features/feature-*/status.json 2>/dev/null | while read f; do cat "$f"; done
+```
+
+Then formats:
+```
+Features
+════════
+Active (2):
+  🔄 user-auth - implementation (60%)
+  🔍 dashboard - analysis (20%)
+
+Paused (1):
+  ⏸ notifications - blocked on API
+
+Completed (3):
+  ✅ login-page, signup-form, forgot-password
+```
+
+### pa:feature-complete
+
+1. Move feature from `active_features` to `completed_features` in `_index.json`
+2. Update feature's `status.json` with completion timestamp
+3. Generate changelog entry
+4. Suggest PR creation if git enabled
+
+### pa:fix "description"
+
+Quick bug fix mode (bypasses full workflow):
+
+1. Analyze the bug description
+2. Search codebase for relevant code
+3. Implement fix directly
+4. Run relevant tests
+5. Log fix in activity.log
+
+---
+
+## Workflow Phase Commands
+
+### pa:analyze
+
+1. Read `./.proagents/prompts/01-analysis.md`
+2. Scan project structure
+3. Identify:
+   - Framework/tech stack
+   - Code patterns
+   - Dependencies
+   - Architecture style
+4. Cache results in `./.proagents/cache/`
+5. Output analysis summary
+
+### pa:requirements
+
+1. Read `./.proagents/prompts/02-requirements.md`
+2. If feature specified, focus on that feature
+3. Document:
+   - Functional requirements
+   - Non-functional requirements
+   - Acceptance criteria
+4. Save to feature folder
+
+### pa:design
+
+1. Read `./.proagents/prompts/03-ui-design.md`
+2. Create:
+   - UI mockups/wireframes (if applicable)
+   - Component structure
+   - Architecture decisions
+3. Document design decisions
+
+### pa:plan
+
+1. Read `./.proagents/prompts/04-planning.md`
+2. Create implementation plan:
+   - Files to create/modify
+   - Order of implementation
+   - Dependencies between tasks
+   - Estimated complexity
+3. Save plan to feature folder
+
+### pa:implement
+
+1. Read `./.proagents/prompts/05-implementation.md`
+2. Follow the plan created in planning phase
+3. Write code following project patterns
+4. Create/update tests as you go
+5. Update progress in status.json
+
+### pa:test-mobile
+
+**AI sets up and runs mobile E2E tests:**
+
+1. Check if Maestro/Detox installed:
+   ```bash
+   maestro --version 2>/dev/null || echo "Not installed"
+   ```
+
+2. If NOT installed - INSTALL IT:
+   ```bash
+   curl -Ls "https://get.maestro.mobile.dev" | bash
+   ```
+
+3. If no test flows exist - CREATE THEM:
+   ```bash
+   mkdir -p .maestro
+   # AI creates test flows based on the feature being tested
+   ```
+
+4. RUN the tests:
+   ```bash
+   maestro test .maestro/
+   ```
+
+5. Report results
+
+**WRONG:** "E2E not configured, test manually..."
+**CORRECT:** "Installing Maestro... Creating tests... Running..."
+
+---
+
+### pa:test
+
+**AI runs tests and shows results:**
+
+```bash
+# AI executes (check config first):
+npm test 2>&1
+# OR from proagents.config.yaml testing.tools.unit.command
+```
+
+Then reports:
+```
+Test Results
+════════════
+✓ 45 passed
+✗ 2 failed
+○ 3 skipped
+
+Failed:
+• src/auth/login.test.ts:23 - Expected true, got false
+• src/api/user.test.ts:45 - Timeout after 5000ms
+```
+
+**NEVER say "run npm test" - actually run it!**
+
+### pa:review
+
+1. Read `./.proagents/prompts/06.5-code-review.md`
+2. Review all changes in current feature
+3. Check:
+   - Code quality
+   - Test coverage
+   - Security issues
+   - Performance concerns
+4. Generate review report
+
+### pa:doc
+
+**AI MUST create actual documentation files, not just show options.**
+
+1. Read `./.proagents/prompts/07-documentation.md`
+2. Analyze codebase structure
+3. **CREATE** documentation files:
+
+```bash
+# AI runs:
+mkdir -p docs docs/api docs/modules docs/components
+
+# Then AI CREATES these files:
+# - ./docs/README.md (project overview)
+# - ./docs/api/*.md (API documentation)
+# - ./docs/modules/*.md (module docs)
+# - ./docs/components/*.md (component docs)
+```
+
+4. Sub-commands:
+   - `pa:doc-api` → Creates `./docs/api/*.md`
+   - `pa:doc-module auth` → Creates `./docs/modules/auth.md`
+   - `pa:doc-file src/api.ts` → Creates doc for that file
+   - `pa:doc-readme` → Updates `./README.md`
+   - `pa:changelog` → Updates `./CHANGELOG.md`
+   - `pa:release` → Creates `./RELEASE_NOTES.md`
+
+5. **Output:**
+```
+Documentation generated:
+✓ ./docs/README.md
+✓ ./docs/api/endpoints.md
+✓ ./docs/modules/auth.md
+...
+Total: 12 files created
+```
+
+### pa:deploy
+
+1. Read `./.proagents/prompts/08-deployment.md`
+2. Run pre-deployment checks
+3. Create deployment checklist
+4. Generate release notes if needed
+
+---
+
+## Documentation Commands
+
+### pa:release
+
+Generate release notes:
+
+1. Read recent commits/changes
+2. Categorize changes (features, fixes, breaking)
+3. Generate formatted release notes
+4. Save to `./RELEASE_NOTES.md`
+
+### pa:changelog
+
+**AI updates ALL changelog files for cross-AI continuity.**
+
+1. Update `./CHANGELOG.md` (public changelog)
+2. Prepend to `.proagents/changelog/_recent.md` (last 10 changes)
+3. Update feature changelog if working on a feature
+4. Update module changelog based on files modified
+
+```bash
+# AI workflow:
+# 1. Identify what changed
+# 2. Determine feature/module from context
+
+# 3. Update root changelog
+# Edit ./CHANGELOG.md
+
+# 4. Prepend to recent changes
+# Edit .proagents/changelog/_recent.md
+
+# 5. Update feature changelog (if applicable)
+# Edit .proagents/changelog/features/[feature-name].md
+
+# 6. Update module changelog (auto-detect from paths)
+# Edit .proagents/changelog/modules/[module].md
+```
+
+### pa:changelog-feature [name]
+
+View/update specific feature changelog:
+```bash
+cat .proagents/changelog/features/[name].md
+```
+
+### pa:changelog-module [name]
+
+View/update specific module changelog:
+```bash
+cat .proagents/changelog/modules/[name].md
+```
+
+---
+
+## Cross-AI Continuity Commands
+
+### pa:sync
+
+**Run FIRST when starting work on any AI platform.**
+
+AI reads context files to understand current state:
+
+```bash
+# AI executes:
+cat .proagents/worklog/_context.md
+cat .proagents/changelog/_recent.md
+ls -t .proagents/worklog/*.md | head -3 | xargs cat
+cat .proagents/active-features/_index.json
+tail -20 .proagents/activity.log
+```
+
+Then reports:
+```
+Project Context Loaded!
+
+Active Work:
+- Feature: user-auth (70% complete)
+- Last: JWT validation added by Claude
+
+Recent Changes:
+1. Added login validation (Claude, Mar 11)
+2. Fixed API endpoint (Gemini, Mar 10)
+
+Pending:
+- [ ] Complete email verification
+- [ ] Add unit tests
+
+Ready to continue. What would you like to work on?
+```
+
+### pa:session-start
+
+Begin a new work session:
+
+```bash
+# AI creates session log
+DATE=$(date '+%Y-%m-%d')
+# Creates: .proagents/worklog/YYYY-MM-DD-[ai]-001.md
+```
+
+### pa:session-end
+
+Finalize session and update all tracking:
+
+```bash
+# AI updates:
+# 1. Session log with summary and next steps
+# 2. .proagents/worklog/_context.md
+# 3. .proagents/changelog/_recent.md
+# 4. Feature/module changelogs
+# 5. activity.log
+```
+
+Output:
+```
+Session Complete!
+- Duration: 45 min
+- Tasks: 3 completed
+- Files: 5 modified
+
+Updated:
+✓ worklog/_context.md
+✓ worklog/2024-03-11-claude-001.md
+✓ changelog/features/user-auth.md
+✓ changelog/_recent.md
+✓ activity.log
+
+Next AI can continue from: [next steps listed]
+```
+
+### pa:resume
+
+**Quick resume for returning AI - loads context and shows next action.**
+
+```bash
+# AI executes:
+cat .proagents/worklog/_context.md
+cat .proagents/changelog/_recent.md
+ls -t .proagents/worklog/*.md 2>/dev/null | head -2 | tail -1 | xargs cat
+tail -10 .proagents/activity.log
+```
+
+Output:
+```
+Resume Context
+══════════════
+Last Session: 2024-03-11 by Claude (opus-4)
+Duration: 45 min
+
+What Was Done:
+- Added JWT validation
+- Fixed login bug
+- Updated API docs
+
+Pending Tasks:
+1. [ ] Complete email verification
+2. [ ] Add unit tests for auth
+
+Files Recently Modified:
+- src/auth/login.ts (2 hours ago)
+- src/auth/validate.ts (2 hours ago)
+
+Suggested Next Action:
+→ Continue with "Complete email verification"
+
+Ready to proceed?
+```
+
+### pa:conflict-check
+
+**Check if files you're about to modify were changed by another AI.**
+
+AI runs before editing any file:
+
+```bash
+# Check recent changes to files
+cat .proagents/changelog/_recent.md | grep -A2 "Files:"
+cat .proagents/worklog/_context.md | grep -A5 "Recent Changes"
+```
+
+Output if conflict detected:
+```
+⚠️ CONFLICT WARNING
+═══════════════════
+File: src/auth/login.ts
+Last modified: 2 hours ago by Gemini (1.5-pro)
+Changes: Added email validation
+
+Options:
+1. Review changes first (recommended)
+2. Proceed with caution
+3. Coordinate with previous AI's work
+
+Your action?
+```
+
+### Validation Reminder
+
+**AI checks if previous session logged changes properly.**
+
+On `pa:sync`, AI also checks:
+
+```bash
+# Compare git changes vs changelog entries
+git diff --name-only HEAD~3 > /tmp/git_changes.txt
+grep -o "src/[^ ]*" .proagents/changelog/_recent.md > /tmp/logged_changes.txt
+
+# If git has changes not in changelog:
+echo "⚠️ Warning: Some changes may not be logged"
+```
+
+Output if missing logs:
+```
+⚠️ LOGGING VALIDATION
+═════════════════════
+Git shows 5 files changed since last session.
+Changelog shows 2 files logged.
+
+Missing from changelog:
+- src/utils/helpers.ts
+- src/api/endpoints.ts
+- tests/auth.test.ts
+
+Previous AI may have forgotten to log. Consider adding entries.
+```
+
+---
+
+## Issue/Fix Linking
+
+**Link changes to issue numbers in changelogs.**
+
+Format for _recent.md:
+```markdown
+### YYYY-MM-DD - Bug Fix
+**Issue:** #123
+**Module:** auth
+**AI:** Claude (opus-4)
+**Files:** src/auth/login.ts (+5, -2)
+**Summary:** Fixed null check in email validation
+**Closes:** #123
+
+---
+```
+
+AI auto-detects issue numbers from:
+- User message: "fix issue #123" or "closes #456"
+- Branch name: `fix/123-login-bug`
+- Commit message: `Fixes #123`
+
+---
+
+## File-Level Lock Tracking
+
+**Track which files each AI is actively editing.**
+
+When AI starts editing a file:
+
+```bash
+# Add to .proagents/.active-files
+echo "src/auth/login.ts|Claude|opus-4|$(date -Iseconds)" >> .proagents/.active-files
+```
+
+Before editing any file, AI checks:
+```bash
+grep "src/auth/login.ts" .proagents/.active-files
+```
+
+If file is locked by another AI:
+```
+⚠️ FILE IN USE
+══════════════
+File: src/auth/login.ts
+Locked by: Gemini (1.5-pro)
+Since: 10 minutes ago
+
+Options:
+1. Wait and retry
+2. Edit different file
+3. Force proceed (may cause conflicts)
+```
+
+On session end, AI clears its file locks:
+```bash
+grep -v "|Claude|" .proagents/.active-files > /tmp/active && mv /tmp/active .proagents/.active-files
+```
+
+---
+
+## Git Commit Integration
+
+**Auto-populate changelog from git commits.**
+
+### pa:changelog --from-git
+
+AI parses recent commits and updates changelog:
+
+```bash
+# Get recent commits
+git log --oneline --since="24 hours ago" --format="%h|%s|%an"
+```
+
+For each commit, AI:
+1. Extracts files changed: `git show --name-only HASH`
+2. Detects module from file paths
+3. Extracts issue numbers from message
+4. Prepends to `_recent.md`
+
+Example auto-generated entry:
+```markdown
+### 2024-03-11 - Commit abc123
+**Module:** auth (auto-detected)
+**Author:** Developer
+**Files:** src/auth/login.ts, src/auth/validate.ts
+**Summary:** Add JWT token validation
+**Commit:** abc123
+**Issues:** #123, #124
+
+---
+```
+
+---
+
+## Quality Commands
+
+### pa:qa
+
+**AI runs ALL checks and reports:**
+
+```bash
+# AI executes each:
+npm run lint 2>&1
+npm test 2>&1
+npm run test:coverage 2>&1
+npm audit 2>&1
+```
+
+Then reports:
+```
+QA Report
+═════════
+Lint: ✓ No issues
+Tests: ✓ 45/45 passed
+Coverage: 82% (target: 80%) ✓
+Security: 0 vulnerabilities ✓
+
+Overall: PASSED
+```
+
+**NEVER say "run these commands" - run them!**
+
+### pa:qa-security
+
+Security-focused audit:
+
+1. Check for common vulnerabilities (OWASP)
+2. Scan dependencies for known issues
+3. Review auth/permissions code
+4. Generate security report
+
+### pa:lint
+
+**AI runs linter and shows results:**
+
+```bash
+# AI executes:
+npm run lint 2>&1
+# If errors, auto-fix:
+npm run lint -- --fix 2>&1
+```
+
+Then reports:
+```
+Lint Results
+════════════
+✓ 120 files checked
+✗ 3 errors found
+⚠ 5 warnings
+
+Auto-fixed: 2 errors
+Remaining: 1 error in src/utils/helper.ts:45
+
+Fixing remaining issue...
+[AI edits the file to fix]
+```
+
+**NEVER say "run npm run lint" - run it!**
+
+---
+
+## Collaboration Commands
+
+### pa:handoff
+
+**AI creates handoff file:**
+
+```bash
+# AI creates .proagents/handoff.md with content
+```
+
+### pa:handoff-read
+
+**AI reads and displays handoff:**
+
+```bash
+# AI executes:
+cat .proagents/handoff.md 2>/dev/null || echo "No handoff notes"
+```
+
+Then displays the content to continue where previous AI left off.
+
+### pa:handoff (write)
+
+Create handoff notes for other AIs:
+
+1. Summarize current work status
+2. List completed items
+3. List in-progress items
+4. Document blockers
+5. Save to `./.proagents/handoff.md`
+
+### pa:feedback "description"
+
+Log feedback for AI learning:
+
+1. Append to `./.proagents/feedback.md`:
+   ```markdown
+   ## [DATE] Feedback
+   **Type:** correction/preference/pattern
+   **Description:** [user's feedback]
+   **Action:** [what should be done differently]
+   ```
+2. Apply learning to current work
+
+### pa:decisions
+
+**AI reads and displays all decisions:**
+
+```bash
+# AI executes:
+cat .proagents/decisions.md 2>/dev/null || echo "No decisions logged"
+```
+
+### pa:errors
+
+**AI reads and displays error history:**
+
+```bash
+# AI executes:
+cat .proagents/errors.md 2>/dev/null || echo "No errors logged"
+```
+
+### pa:context
+
+**AI reads project context:**
+
+```bash
+# AI executes:
+cat .proagents/context.md 2>/dev/null
+```
+
+This should be read at START of every session!
+
+### pa:feedback
+
+**AI reads feedback/corrections:**
+
+```bash
+# AI executes:
+cat .proagents/feedback.md 2>/dev/null || echo "No feedback"
+```
+
+Learn from these to avoid repeating mistakes!
+
+### pa:decision "title"
+
+Log architectural decision:
+
+1. Create ADR in `./.proagents/adr/`
+2. Document:
+   - Context
+   - Decision
+   - Consequences
+3. Reference in current feature
+
+### pa:activity
+
+**AI runs and displays:**
+
+```bash
+# AI executes:
+cat .proagents/activity.log | grep -v "^#" | tail -20
+```
+
+Then formats output:
+```
+Recent Activity
+═══════════════
+[2024-03-06 16:00] [Gemini] pa:test - 12 tests passed
+[2024-03-06 15:45] [Claude] pa:implement - Created UserService
+```
+
+---
+
+## Configuration Commands
+
+### pa:config
+
+**AI reads and displays config:**
+
+```bash
+# AI executes:
+cat proagents.config.yaml
+```
+
+Then formats:
+```
+ProAgents Configuration
+═══════════════════════
+Project: my-app (nextjs)
+
+Platforms: claude, cursor, copilot
+
+Testing:
+  Framework: vitest
+  Coverage: 80%
+
+Git:
+  Enabled: true
+  Branch prefix: feature/
+```
+
+### pa:checkpoint
+
+Pause for user approval:
+
+1. Summarize completed work
+2. Show next steps
+3. Wait for user confirmation
+4. Log checkpoint in activity
+
+### pa:skip-checkpoint
+
+Skip the current checkpoint and continue.
+
+---
+
+## History & Progress Commands
+
+### pa:history
+
+**AI MUST read the actual activity.log file and display contents.**
+
+1. Run: `cat .proagents/activity.log | tail -30`
+2. Parse and format entries
+3. Display most recent first
+4. NEVER say "no commands yet" without reading the file first
+
+**Wrong behavior:**
+```
+"No commands recorded yet"  ← Without reading file!
+```
+
+**Correct behavior:**
+```
+Reading .proagents/activity.log...
+
+Command History
+═══════════════
+[2024-03-06 16:00] [Gemini] pa:logs - Captured 50 entries
+[2024-03-06 15:30] [Cursor] pa:test - 12 tests passed
+```
+
+### pa:progress
+
+**AI MUST read feature status files and calculate progress.**
+
+1. Run: `cat .proagents/active-features/_index.json`
+2. Read each feature's status.json
+3. Calculate progress percentage
+4. Show visual progress bar
+
+### pa:activity
+
+**AI MUST read and display activity.log:**
+
+1. Run: `cat .proagents/activity.log | tail -20`
+2. Show recent AI activity
+3. Highlight actions by different AI platforms
+
+---
+
+## Session Commands
+
+### pa:session-end
+
+Generate session summary:
+
+1. List all changes made
+2. List files modified
+3. List commands executed
+4. Document any issues
+5. Save to session history
+
+### pa:lock
+
+**AI reads and displays lock status:**
+
+```bash
+# AI executes:
+cat .proagents/.lock 2>/dev/null || echo "No lock file"
+```
+
+Then reports:
+```
+Lock Status
+═══════════
+Locked by: Claude (opus-4)
+Task: pa:feature user-auth
+Started: 2024-03-06 15:00
+Expires: 2024-03-06 17:00
+
+Or if no lock:
+No active lock. Safe to proceed.
+```
+
+### pa:lock-release
+
+Release lock if you hold it:
+
+1. Verify you hold the lock
+2. Delete `./.proagents/.lock`
+3. Log release in activity
+
+---
+
+## Debug & Log Commands
+
+**CRITICAL: AI must RUN log commands itself, never tell user to "check the logs".**
+
+### pa:debug
+
+Start a debug session:
+
+1. Detect project platform (Web, React Native, Android, iOS)
+2. Identify the issue from user description
+3. **RUN log capture commands:**
+   - React Native: `adb logcat -d *:S ReactNativeJS:V | tail -200`
+   - Android: `adb logcat -d -s AppTag:D | tail -200`
+   - iOS: `xcrun simctl spawn booted log show --last 5m`
+4. **ANALYZE** the captured output
+5. **REPORT** findings and implement fix
+
+### pa:debug-add
+
+Add debug logs throughout code:
+
+1. Detect platform (Web/RN/Android/iOS)
+2. Analyze code structure
+3. Add appropriate logging:
+   - Function entry/exit with params
+   - Variable state changes
+   - API calls and responses
+   - Error conditions
+4. Mark all logs with `// DEBUG:START` and `// DEBUG:END`
+5. Use platform-appropriate syntax:
+   - JS/TS: `console.log('[DEBUG]', ...)`
+   - Android: `Log.d("DEBUG", ...)`
+   - iOS: `print("[DEBUG]", ...)`
+
+### pa:debug-trace "function"
+
+Add detailed tracing to specific function:
+
+1. Find function in codebase
+2. Add entry log with all parameters
+3. Add timing measurement
+4. Add exit log with return value
+5. Wrap in try-catch with error logging
+
+### pa:debug-var "variable"
+
+Track all changes to a variable:
+
+1. Find variable declaration
+2. Add change tracking (willSet/didSet, proxy, setter)
+3. Log old value → new value
+4. Include stack trace for debugging
+
+### pa:debug-api
+
+Add API request/response logging:
+
+1. Find all API/fetch calls
+2. Add request logging (method, url, body)
+3. Add response logging (status, data, timing)
+4. Add error logging with full details
+
+### pa:debug-state
+
+Add state change logging:
+
+1. Find state management code
+2. Add logging for state changes:
+   - React: useState, useReducer, Redux, Zustand
+   - Android: ViewModel, LiveData, StateFlow
+   - iOS: @State, @Published, ObservableObject
+3. Log before/after values
+
+### pa:debug-error
+
+Add comprehensive error logging:
+
+1. Wrap risky code in try-catch
+2. Add error boundaries (React)
+3. Log error message, stack, context
+4. Add breadcrumb trail
+
+### pa:debug-web
+
+Web/browser debugging:
+
+1. Check browser console for errors
+2. Inspect network requests
+3. Review console.log statements
+4. Check for debugger statements
+5. Analyze stack traces
+
+### pa:debug-rn
+
+React Native debugging:
+
+1. Check Metro bundler logs
+2. Use Flipper or Reactotron
+3. View device logs via ADB (Android) or Console.app (iOS)
+4. Check `__DEV__` conditional logs
+5. Analyze red screen errors
+
+### pa:debug-android
+
+Android native debugging:
+
+1. Run `adb logcat` with appropriate filters
+2. Filter by tag: `adb logcat -s MyApp:D`
+3. Filter by priority: `adb logcat *:E` (errors only)
+4. Check for crashes in Logcat
+5. Use Android Studio Logcat panel
+
+### pa:debug-ios
+
+iOS native debugging:
+
+1. Use Xcode console (Cmd+Shift+C)
+2. Check Console.app with device selected
+3. Filter by subsystem/category with os_log
+4. Run `xcrun simctl spawn booted log stream`
+5. Check crash logs in Organizer
+
+### pa:logs
+
+**AI runs log commands and shows output:**
+
+1. Detect platform
+2. Run appropriate command:
+   - RN: `adb logcat -d *:S ReactNativeJS:V | tail -100`
+   - Android: `adb logcat -d | tail -100`
+   - iOS: `xcrun simctl spawn booted log show --last 2m`
+3. Parse and display results
+4. Highlight errors and warnings
+
+### pa:logs-filter "term"
+
+**AI runs filtered log command:**
+
+1. Run: `adb logcat -d | grep -i "term"` (or platform equivalent)
+2. Parse output
+3. Show matching entries with context
+4. Analyze patterns in matches
+
+### pa:logs-clear
+
+Clear debug logs:
+
+- Web: Clear browser console
+- React Native: Restart Metro bundler
+- Android: `adb logcat -c`
+- iOS: Clear in Console.app
+
+### pa:debug-clean
+
+Remove debug statements before production:
+
+1. Find all debug statements:
+   - `console.log`, `console.debug`, `debugger` (JS)
+   - `Log.d`, `Log.v`, `System.out` (Android)
+   - `print`, `debugPrint`, `NSLog` (iOS)
+2. List found statements with file:line
+3. Offer to remove or comment out
+4. Verify no production-breaking changes