proagents 1.6.4 → 1.6.6

package/.proagents/AI_INSTRUCTIONS.md

@@ -273,6 +273,49 @@ When the user types commands starting with `pa:`, recognize and execute them.
 | `pa:help` | Show all available commands |
 | `pa:status` | Show current progress |
 
+**How to execute Initialization commands:**
+
+For `pa:status` (or `pa:s`):
+1. **Read the ACTUAL data** from `./.proagents/active-features/_index.json`
+2. **Check if empty** - if `active_features`, `paused_features`, and `completed_features` are all empty arrays:
+   ```
+   Project Status
+   ══════════════
+
+   No features currently tracked.
+
+   Start a new feature with:
+   → pa:feature-start "feature name"
+
+   Or quick bug fix:
+   → pa:fix "bug description"
+   ```
+
+3. **If has features**, show real status:
+   ```
+   Project Status
+   ══════════════
+
+   Active Features: [count]
+   ────────────────────────
+   • [feature-name] - [phase] ([progress]%)
+     Branch: [branch]
+     Last updated: [timestamp]
+
+   Paused Features: [count]
+   ────────────────────────
+   • [feature-name] - Paused: [reason]
+
+   Completed Features: [count]
+   ──────────────────────────
+   • [feature-name] - Completed [date]
+
+   Next Steps:
+   → [suggested action based on current state]
+   ```
+
+4. **NEVER show example/template data** - only show what's actually in the JSON file
+
 ### Feature Development
 | Command | Action |
 |---------|--------|
@@ -361,6 +404,119 @@ For `pa:implement`:
 | `pa:deploy` | Deployment preparation |
 | `pa:rollback` | Rollback procedures |
 
+**How to execute Quality, Testing & Deployment commands (FULL AUTOMATION):**
+
+For `pa:test`:
+1. **Detect project type** and test framework (Jest, Vitest, Mocha, Pytest, etc.)
+2. **Auto-install missing test dependencies** (no confirmation)
+3. **Run all tests:**
+   ```
+   Running tests...
+   ✓ 45 passed
+   ✗ 3 failed
+   ```
+4. **Auto-fix ALL failures** (no confirmation):
+   ```
+   AUTO-FIX: Fixing 3 failed tests...
+
+   Fix 1/3: UserService.test.ts
+   ─────────────────────────────
+   Root cause: Expected 'user' but got 'undefined'
+   Fixing: src/services/UserService.ts:23
+   → Added null check before return
+   Re-running test... ✓ PASSED
+   ```
+5. **Loop until all pass** or fix is impossible
+6. **Generate report** at `.proagents/test-reports/`
+
+For `pa:qa`:
+1. **Run ALL quality checks** automatically:
+   - Linting (ESLint, Prettier)
+   - Type checking (TypeScript)
+   - Security scan (npm audit)
+   - Dead code detection
+2. **Auto-fix ALL issues** (no confirmation):
+   ```
+   Running QA checks...
+
+   ESLint: 12 errors found
+   AUTO-FIX: Fixing...
+   → Fixed 12/12 errors
+
+   TypeScript: 3 type errors
+   AUTO-FIX: Fixing...
+   → src/api/user.ts:15 - Added type annotation
+   → src/utils/date.ts:8 - Fixed return type
+   → src/hooks/useAuth.ts:22 - Added null check
+
+   Prettier: 28 files need formatting
+   AUTO-FIX: Formatting... ✓ Done
+
+   npm audit: 2 vulnerabilities
+   AUTO-FIX: Running npm audit fix... ✓ Fixed
+   ```
+3. **Re-run checks** to verify all fixed
+4. **Only report issues that cannot be auto-fixed**
+
+For `pa:fix "description"`:
+1. **Understand the bug** from description
+2. **Find the root cause** automatically
+3. **Apply the fix** directly (no confirmation)
+4. **Run related tests** to verify fix works
+5. **Auto-fix any test failures** caused by the fix
+6. **Commit when done** (if git is configured)
+   ```
+   pa:fix "login button not working"
+
+   Finding root cause...
+   → Found: onClick handler missing in LoginButton.tsx:34
+
+   Applying fix...
+   → Added: onClick={() => handleLogin()}
+
+   Running related tests...
+   ✓ LoginButton.test.tsx - 5/5 passed
+
+   ✓ Fix complete
+   ```
+
+For `pa:deploy`:
+1. **Run pre-deployment checks** automatically:
+   - All tests passing? If not → auto-fix
+   - QA checks passing? If not → auto-fix
+   - Security scan clean? If not → auto-fix
+   - Build succeeds? If not → auto-fix
+2. **Only proceed when all checks pass**
+3. **Deploy to target environment**
+4. **Run post-deployment verification**
+5. **Auto-rollback if verification fails**
+   ```
+   pa:deploy
+
+   Pre-deployment checks...
+   ✓ Tests: 128/128 passed
+   ✓ QA: All checks passed
+   ✓ Security: No vulnerabilities
+   ✓ Build: Successful
+
+   Deploying to staging...
+   ✓ Deployed successfully
+
+   Post-deployment verification...
+   ✓ Health check passed
+   ✓ Smoke tests passed
+
+   ══════════════════════════════════════════
+   DEPLOYMENT SUCCESSFUL ✓
+   ══════════════════════════════════════════
+   ```
+
+**IMPORTANT - Full Automation Rules:**
+- **Never ask** "Should I fix this?" - Just fix it
+- **Never say** "Run `pa:test` to verify" - Run it yourself
+- **Never give** "Next steps" - Do the steps
+- **Only stop when:** All done OR truly impossible to proceed
+
 ### Navigation & Flow
 | Command | Action |
 |---------|--------|
@@ -1999,45 +2155,97 @@ For `pa:alias-add`:
    Runs: pa:test && pa:build && pa:deploy
    ```
 
-### Mobile Test Suite (React Native / Mobile Apps)
+### Platform-Specific Test Suites
+
+**Mobile (React Native / Expo):**
 | Command | Action |
 |---------|--------|
 | `pa:test-mobile` | Run full mobile test suite |
 | `pa:test-mobile "feature"` | Test specific feature |
+
+**Web (React / Next.js / Vue / etc.):**
+| Command | Action |
+|---------|--------|
+| `pa:test-web` | Run full web test suite |
+| `pa:test-web "feature"` | Test specific web feature |
+
+**API / Backend (Node.js / Python / etc.):**
+| Command | Action |
+|---------|--------|
+| `pa:test-api` | Run full API/backend test suite |
+| `pa:test-api "endpoint"` | Test specific API endpoint |
+
+**Visual & Comparison:**
+| Command | Action |
+|---------|--------|
 | `pa:test-visual` | Visual/design comparison testing |
-| `pa:test-auto-fix` | Auto-fix failing tests |
-| `pa:test-loop` | Test → Fix → Retest loop until success |
 | `pa:compare-figma` | Compare UI against Figma design |
 | `pa:compare-image "path"` | Compare UI against image/sketch |
 | `pa:screenshot` | Take app screenshots for comparison |
 
+**Auto-Fix:**
+| Command | Action |
+|---------|--------|
+| `pa:test-auto-fix` | Auto-fix failing tests |
+| `pa:test-loop` | Test → Fix → Retest loop until success |
+
 **How to execute Mobile Test Suite commands:**
 
 For `pa:test-mobile`:
-1. **FIRST: Check required tools are installed (FAIL if missing, do NOT skip):**
+
+**Supports ALL mobile platforms:**
+- React Native / Expo
+- Native Android (Kotlin/Java)
+- Native iOS (Swift/Objective-C)
+- Flutter (Dart)
+
+1. **Detect project type automatically:**
    ```
-   Checking required tools...
-   ✓ Node.js (v18.0.0)
-   ✓ Jest / Vitest
-   ✓ React Native Testing Library
-   ✗ MISSING: Maestro - Required for E2E tests
-   ✗ MISSING: Detox - Required for E2E tests
+   Detecting mobile project type...
 
-   ══════════════════════════════════════════════════════════
-   ERROR: Missing required tools. Cannot run mobile test suite.
-   ══════════════════════════════════════════════════════════
+   Checking for:
+   ├── package.json + react-native → React Native
+   ├── android/build.gradle (no RN) → Native Android
+   ├── ios/*.xcodeproj (no RN) → Native iOS
+   ├── pubspec.yaml → Flutter
 
-   Missing tools detected. Auto-installing...
+   Detected: [PROJECT_TYPE]
+   ```
+
+2. **Check & auto-install required tools based on platform:**
+
+   **React Native / Expo:**
+   ```
+   ✓ Node.js
+   ✓ Jest / Vitest
+   ✓ React Native Testing Library
+   ✗ MISSING: Maestro → Auto-installing...
+   ✗ MISSING: Detox → Auto-installing...
+   ```
 
-   Installing Maestro...
-   → curl -Ls "https://get.maestro.mobile.dev" | bash
-   ✓ Maestro installed successfully
+   **Native Android (Kotlin/Java):**
+   ```
+   ✓ Android Studio / Gradle
+   ✓ JUnit
+   ✗ MISSING: Espresso → Auto-adding to build.gradle
+   ✗ MISSING: UI Automator → Auto-adding to build.gradle
+   ✗ MISSING: Maestro → Auto-installing...
+   ```
 
-   Installing Detox CLI...
-   → npm install -g detox-cli
-   ✓ Detox CLI installed successfully
+   **Native iOS (Swift/Objective-C):**
+   ```
+   ✓ Xcode
+   ✓ XCTest (built-in)
+   ✓ XCUITest (built-in)
+   ✗ MISSING: Maestro → Auto-installing...
+   ```
 
-   All tools installed. Running test suite...
+   **Flutter:**
+   ```
+   ✓ Flutter SDK
+   ✓ flutter_test (built-in)
+   ✗ MISSING: integration_test → Auto-adding to pubspec.yaml
+   ✗ MISSING: Maestro → Auto-installing...
    ```
 
    **IMPORTANT - FULL AUTOMATION:**
@@ -2046,15 +2254,11 @@ For `pa:test-mobile`:
    - Fix doesn't work? **Try alternative fix** (no confirmation)
    - Only stop when: all tests pass OR no fix possible
 
-   **Do NOT:**
-   - Ask "Should I install missing tools?"
-   - Ask "Should I fix this?"
-   - Say "Run `pa:test-auto-fix` to fix"
-   - Give "Next steps" for user to follow
-
    **Just DO it.**
 
-2. If all tools present, run comprehensive mobile test suite:
+3. **Run platform-specific test suite:**
+
+   **For React Native:**
    ```
    Mobile Test Suite
    ═════════════════
@@ -2145,6 +2349,129 @@ For `pa:test-mobile`:
    ══════════════════════════════════════════
    ```
 
+   **For Native Android (Kotlin/Java):**
+   ```
+   Android Test Suite
+   ══════════════════
+
+   Phase 1: Unit Tests (JUnit)
+   ───────────────────────────
+   Running: ./gradlew test
+   ✓ UserRepositoryTest - 8 passed
+   ✓ AuthViewModelTest - 12 passed
+   ✓ PaymentServiceTest - 6 passed
+
+   Phase 2: Instrumented Tests (Espresso)
+   ──────────────────────────────────────
+   Running: ./gradlew connectedAndroidTest
+   ✓ LoginActivityTest - UI renders correctly
+   ✓ MainActivityTest - Navigation works
+   ✗ FAILED: CheckoutActivityTest - Button not found
+
+   AUTO-FIX: Analyzing failure...
+   Root cause: Button ID changed
+   Fixing: app/src/main/res/layout/activity_checkout.xml
+   → Changed: android:id="@+id/btn_pay" → android:id="@+id/checkout_button"
+   Fixing: app/src/androidTest/.../CheckoutActivityTest.kt
+   → Changed: R.id.btn_pay → R.id.checkout_button
+   Re-running test... ✓ PASSED
+
+   Phase 3: UI Tests (UI Automator)
+   ─────────────────────────────────
+   ✓ Full checkout flow
+   ✓ User settings update
+   ✓ Push notification handling
+
+   Phase 4: E2E Tests (Maestro)
+   ────────────────────────────
+   Running: maestro test .maestro/
+   ✓ User login flow
+   ✓ Product purchase flow
+
+   ══════════════════════════════════════════
+   ALL TESTS PASSING: 32/32 ✓
+   ══════════════════════════════════════════
+   ```
+
+   **For Native iOS (Swift/Objective-C):**
+   ```
+   iOS Test Suite
+   ══════════════
+
+   Phase 1: Unit Tests (XCTest)
+   ────────────────────────────
+   Running: xcodebuild test -scheme MyApp -destination 'platform=iOS Simulator,name=iPhone 15'
+   ✓ UserServiceTests - 10 passed
+   ✓ AuthManagerTests - 8 passed
+   ✓ NetworkClientTests - 12 passed
+
+   Phase 2: UI Tests (XCUITest)
+   ────────────────────────────
+   ✓ LoginViewControllerTests - UI renders
+   ✓ HomeViewControllerTests - Navigation works
+   ✗ FAILED: ProfileViewControllerTests - Label not found
+
+   AUTO-FIX: Analyzing failure...
+   Root cause: Accessibility identifier missing
+   Fixing: ProfileViewController.swift:45
+   → Added: nameLabel.accessibilityIdentifier = "profile_name_label"
+   Re-running test... ✓ PASSED
+
+   Phase 3: E2E Tests (Maestro)
+   ────────────────────────────
+   Running: maestro test .maestro/
+   ✓ Onboarding flow
+   ✓ Purchase flow
+   ✓ Settings update
+
+   ══════════════════════════════════════════
+   ALL TESTS PASSING: 28/28 ✓
+   ══════════════════════════════════════════
+   ```
+
+   **For Flutter (Dart):**
+   ```
+   Flutter Test Suite
+   ══════════════════
+
+   Phase 1: Unit Tests
+   ───────────────────
+   Running: flutter test test/unit/
+   ✓ user_repository_test.dart - 8 passed
+   ✓ auth_bloc_test.dart - 15 passed
+   ✓ api_client_test.dart - 10 passed
+
+   Phase 2: Widget Tests
+   ─────────────────────
+   Running: flutter test test/widget/
+   ✓ login_screen_test.dart - renders correctly
+   ✓ home_screen_test.dart - shows user data
+   ✗ FAILED: cart_screen_test.dart - widget not found
+
+   AUTO-FIX: Analyzing failure...
+   Root cause: Key not set on widget
+   Fixing: lib/screens/cart_screen.dart:67
+   → Added: key: const Key('checkout_button')
+   Re-running test... ✓ PASSED
+
+   Phase 3: Integration Tests
+   ──────────────────────────
+   Running: flutter test integration_test/
+   ✓ app_test.dart - Full app flow
+   ✓ checkout_test.dart - Purchase complete
+
+   Phase 4: E2E Tests (Maestro)
+   ────────────────────────────
+   Running: maestro test .maestro/
+   ✓ User registration
+   ✓ Product browsing
+   ✓ Checkout flow
+
+   ══════════════════════════════════════════
+   ALL TESTS PASSING: 42/42 ✓
+   ══════════════════════════════════════════
+   ```
+
 4. If a fix cannot be applied automatically:
    ```
    ✗ Cannot auto-fix: Database schema mismatch
@@ -2156,6 +2483,167 @@ For `pa:test-mobile`:
 
 5. Generate detailed test report at `./.proagents/test-reports/mobile-{timestamp}.md`
 
+For `pa:test-web`:
+1. **Check required tools and auto-install:**
+   ```
+   Checking required tools...
+   ✓ Node.js
+   ✓ Vitest / Jest
+   ✗ MISSING: Playwright
+
+   Auto-installing Playwright...
+   → npx playwright install
+   ✓ Playwright installed
+   ```
+
+2. **Run full web test suite:**
+   ```
+   Web Test Suite
+   ══════════════
+
+   Phase 1: Unit Tests (Vitest/Jest)
+   ──────────────────────────────────
+   Running: npm test
+   ✓ utils/formatDate.test.ts - 5 passed
+   ✓ hooks/useAuth.test.ts - 8 passed
+   ✓ services/api.test.ts - 12 passed
+
+   Phase 2: Component Tests
+   ────────────────────────
+   ✓ Button.test.tsx - renders correctly
+   ✓ Form.test.tsx - validates inputs
+   ✗ FAILED: Modal.test.tsx - close handler
+
+   AUTO-FIX: Analyzing failure...
+   Root cause: onClose prop not called
+   Fixing: src/components/Modal.tsx:24
+   → Added: onClick={() => onClose?.()}
+   Re-running test... ✓ PASSED
+
+   Phase 3: Integration Tests
+   ──────────────────────────
+   ✓ Login flow
+   ✓ Checkout process
+   ✓ User settings update
+
+   Phase 4: E2E Tests (Playwright)
+   ────────────────────────────────
+   Running: npx playwright test
+   ✓ Homepage loads correctly
+   ✓ User can login
+   ✓ User can add to cart
+   ✗ FAILED: Checkout completes
+
+   AUTO-FIX: Analyzing failure...
+   Root cause: Submit button selector changed
+   Fixing: e2e/checkout.spec.ts:45
+   → Changed: 'button.submit' → 'button[type="submit"]'
+   Re-running test... ✓ PASSED
+
+   Phase 5: Visual Regression (Percy/Playwright)
+   ──────────────────────────────────────────────
+   ✓ Homepage - no visual changes
+   ✓ Product page - no visual changes
+   ✗ FAILED: Cart page - button color changed
+
+   AUTO-FIX: Analyzing failure...
+   Root cause: CSS variable override
+   Fixing: src/styles/cart.css:12
+   → Changed: --btn-color: blue → --btn-color: var(--primary)
+   Re-running test... ✓ PASSED
+
+   ══════════════════════════════════════════
+   ALL TESTS PASSING: 45/45 ✓
+   ══════════════════════════════════════════
+   ```
+
+3. Generate detailed test report at `./.proagents/test-reports/web-{timestamp}.md`
+
+For `pa:test-api`:
+1. **Check required tools and auto-install:**
+   ```
+   Checking required tools...
+   ✓ Node.js / Python
+   ✓ Jest / Pytest
+   ✗ MISSING: Supertest
+   ✗ MISSING: k6 (load testing)
+
+   Auto-installing...
+   → npm install --save-dev supertest
+   → brew install k6
+   ✓ All tools installed
+   ```
+
+2. **Run full API test suite:**
+   ```
+   API Test Suite
+   ══════════════
+
+   Phase 1: Unit Tests
+   ───────────────────
+   ✓ UserService - 12 tests passed
+   ✓ AuthService - 8 tests passed
+   ✓ PaymentService - 15 tests passed
+
+   Phase 2: Integration Tests
+   ──────────────────────────
+   Testing: GET /api/users
+   ✓ Returns 200 with user list
+   ✓ Supports pagination
+   ✓ Filters by role
+
+   Testing: POST /api/auth/login
+   ✓ Returns token on valid credentials
+   ✓ Returns 401 on invalid password
+   ✗ FAILED: Returns 429 on rate limit
+
+   AUTO-FIX: Analyzing failure...
+   Root cause: Rate limiter not configured in test env
+   Fixing: src/middleware/rateLimit.ts:8
+   → Added: if (process.env.NODE_ENV === 'test') return next()
+   Re-running test... ✓ PASSED
+
+   Phase 3: Contract Tests
+   ───────────────────────
+   Validating OpenAPI spec...
+   ✓ GET /api/users matches schema
+   ✓ POST /api/users matches schema
+   ✗ FAILED: PUT /api/users/:id - missing field
+
+   AUTO-FIX: Analyzing failure...
+   Root cause: Response missing 'updatedAt' field
+   Fixing: src/controllers/userController.ts:89
+   → Added: updatedAt: user.updatedAt to response
+   Re-running test... ✓ PASSED
+
+   Phase 4: Load Tests (k6)
+   ────────────────────────
+   Running: k6 run loadtest.js
+   ✓ 100 VUs, 30s duration
+   ✓ p95 response time: 145ms (< 200ms threshold)
+   ✓ Error rate: 0.1% (< 1% threshold)
+   ✓ Throughput: 850 req/s
+
+   Phase 5: Security Tests
+   ───────────────────────
+   ✓ SQL injection: Protected
+   ✓ XSS: Protected
+   ✓ CSRF: Protected
+   ✗ FAILED: Rate limiting on /api/auth/login
+
+   AUTO-FIX: Analyzing failure...
+   Root cause: Rate limit too high (1000 req/min)
+   Fixing: src/config/security.ts:15
+   → Changed: loginRateLimit: 1000 → loginRateLimit: 10
+   Re-running test... ✓ PASSED
+
+   ══════════════════════════════════════════
+   ALL TESTS PASSING: 52/52 ✓
+   ══════════════════════════════════════════
+   ```
+
+3. Generate detailed test report at `./.proagents/test-reports/api-{timestamp}.md`
+
 For `pa:test-visual`:
 1. Take screenshots of all screens/components
 2. Compare against design references:
@@ -2261,6 +2749,65 @@ For `pa:test-auto-fix`:
    Running verification tests...
    ```
 
+4. **Learn from fixes** - Store successful fixes for future use:
+   ```
+   Saving fix patterns to .proagents/.learning/fixes.json...
+
+   New patterns learned:
+   ┌────────────────────┬────────────────────────────────────────┐
+   │ Error Pattern      │ Fix Pattern                            │
+   ├────────────────────┼────────────────────────────────────────┤
+   │ "undefined" error  │ Add null check before access           │
+   │ testID not found   │ Add testID prop to component           │
+   │ color mismatch     │ Use design token instead of hex        │
+   │ rate limit fail    │ Disable in test environment            │
+   │ schema mismatch    │ Add missing field to response          │
+   └────────────────────┴────────────────────────────────────────┘
+
+   ✓ 5 fix patterns saved
+   ```
+
+**Auto-Fix Intelligence:**
+
+The AI learns from previous fixes to apply them faster:
+
+1. **Pattern Database** - Stored in `.proagents/.learning/`:
+   ```
+   .proagents/.learning/
+   ├── fixes.json           # Successful fix patterns
+   ├── errors.json          # Common error patterns
+   └── project-patterns.json # Project-specific patterns
+   ```
+
+2. **Fix Matching** - When a test fails:
+   ```
+   Analyzing failure: "Cannot read property 'id' of undefined"
+
+   Checking learned patterns...
+   ✓ Match found: "undefined property access"
+
+   Applying known fix:
+   → Add optional chaining: user?.id
+
+   Confidence: 95% (applied 12 times before, 100% success rate)
+   ```
+
+3. **Learning from Corrections** - If AI fix is wrong:
+   ```
+   User corrected fix:
+   - AI suggested: user?.id ?? 'default'
+   - User changed to: user?.id || throw new Error('User required')
+
+   Learning...
+   → Updated pattern: In auth contexts, throw error instead of default
+   → Saved to .proagents/.learning/corrections.json
+   ```
+
+4. **Project-Specific Learning**:
+   - Learns project conventions (e.g., "this project uses Zod for validation")
+   - Remembers file structure patterns
+   - Knows which solutions worked in this codebase
+
 For `pa:test-loop`:
 1. Run complete Test → Fix → Retest cycle:
    ```
@@ -2512,6 +3059,213 @@ mobile_testing:
 └── auto-fix-log.md                # Log of all auto-fixes
 ```
 
+### Custom Testing Tools Configuration
+
+Users can configure their own testing tools in `proagents.config.yaml`. **AI MUST check and use these custom tools** instead of defaults.
+
+**Read custom tools from config:**
+```yaml
+# proagents.config.yaml → testing.tools section
+testing:
+  tools:
+    unit:
+      command: "npm test"           # Custom unit test command
+      framework: "jest"             # jest | vitest | mocha | pytest | junit | xctest
+    integration:
+      command: "npm run test:integration"
+      framework: "jest"
+    e2e:
+      command: "npx playwright test"
+      framework: "playwright"       # playwright | cypress | maestro | detox
+    visual:
+      command: "npx percy exec -- npm test"
+      framework: "percy"            # percy | chromatic | applitools
+    load:
+      command: "k6 run loadtest.js"
+      framework: "k6"               # k6 | artillery | locust | jmeter
+    security:
+      command: "npm audit && snyk test"
+      framework: "snyk"             # snyk | npm-audit | owasp-zap
+
+  auto_install: true   # Auto-install missing tools
+
+  custom_commands:     # User shortcuts
+    "test:quick": "npm test -- --onlyChanged"
+    "test:ci": "npm test -- --coverage --ci"
+```
+
+**How AI uses custom tools:**
+
+For `pa:test`, `pa:test-mobile`, `pa:test-web`, `pa:test-api`:
+
+1. **Read proagents.config.yaml first:**
+   ```
+   Loading test configuration...
+
+   Custom tools configured:
+   ├── Unit: npm test (jest)
+   ├── Integration: npm run test:integration (jest)
+   ├── E2E: npx playwright test (playwright)
+   ├── Visual: npx percy exec -- npm test (percy)
+   ├── Load: k6 run loadtest.js (k6)
+   └── Security: npm audit && snyk test (snyk)
+   ```
+
+2. **Use custom command instead of default:**
+   ```
+   Running Unit Tests
+   ══════════════════
+
+   Using custom command: npm test
+   Framework: jest
+
+   > npm test
+
+   PASS  src/__tests__/auth.test.ts
+   PASS  src/__tests__/user.test.ts
+   ...
+   ```
+
+3. **Auto-install if tool missing and auto_install: true:**
+   ```
+   Tool not found: k6
+   Config: auto_install: true
+
+   Installing k6...
+   ✓ k6 installed successfully
+
+   Running load tests...
+   ```
+
+4. **Support custom_commands shortcuts:**
+   ```
+   User runs: pa:test quick
+
+   Checking custom_commands...
+   Found: "test:quick" → "npm test -- --onlyChanged"
+
+   Running: npm test -- --onlyChanged
+   ```
+
+**Mobile Testing Tools (platform-specific):**
+
+For `pa:test-mobile`, read the `testing.mobile` section:
+
+```yaml
+# proagents.config.yaml → testing.mobile section
+testing:
+  mobile:
+    # React Native / Expo
+    react_native:
+      unit:
+        command: "npm test"
+        framework: "jest"
+      component:
+        command: "npm test -- --testPathPattern=components"
+        framework: "@testing-library/react-native"
+      e2e:
+        command: "maestro test .maestro/"
+        framework: "maestro"          # maestro | detox | appium
+
+    # Native Android (Kotlin/Java)
+    android:
+      unit:
+        command: "./gradlew test"
+        framework: "junit"
+      integration:
+        command: "./gradlew connectedAndroidTest"
+        framework: "espresso"
+      e2e:
+        command: "maestro test .maestro/"
+        framework: "maestro"
+
+    # Native iOS (Swift/Objective-C)
+    ios:
+      unit:
+        command: "xcodebuild test -scheme MyApp -destination 'platform=iOS Simulator,name=iPhone 15'"
+        framework: "xctest"
+      e2e:
+        command: "maestro test .maestro/"
+        framework: "maestro"
+
+    # Flutter
+    flutter:
+      unit:
+        command: "flutter test"
+        framework: "flutter_test"
+      integration:
+        command: "flutter test integration_test/"
+        framework: "integration_test"
+      e2e:
+        command: "maestro test .maestro/"
+        framework: "maestro"
+```
+
+**How AI uses mobile tools:**
+
+1. **Detect platform first:**
+   ```
+   Detecting mobile project type...
+
+   Checking for:
+   ├── package.json + react-native → React Native
+   ├── android/build.gradle (no RN) → Native Android
+   ├── ios/*.xcodeproj (no RN) → Native iOS
+   ├── pubspec.yaml → Flutter
+
+   Detected: React Native
+   ```
+
+2. **Read platform-specific config:**
+   ```
+   Loading config: testing.mobile.react_native
+
+   Tools:
+   ├── Unit: npm test (jest)
+   ├── Component: npm test -- --testPathPattern=components
+   └── E2E: maestro test .maestro/ (maestro)
+   ```
+
+3. **Run with configured commands:**
+   ```
+   Running React Native Test Suite
+   ═══════════════════════════════
+
+   Phase 1: Unit Tests
+   > npm test
+   ✓ 45 tests passed
+
+   Phase 2: Component Tests
+   > npm test -- --testPathPattern=components
+   ✓ 23 tests passed
+
+   Phase 3: E2E Tests
+   > maestro test .maestro/
+   ✓ 12 flows passed
+   ```
+
+**Default fallbacks (if no custom config):**
+
+| Test Type | Detect From | Default Command |
+|-----------|-------------|-----------------|
+| Unit | package.json scripts | `npm test` or `npx jest` |
+| Integration | test files | `npm run test:integration` |
+| E2E | playwright.config / cypress.json | `npx playwright test` or `npx cypress run` |
+| Visual | percy.yml / chromatic | `npx percy exec` |
+| Load | k6 / artillery config | `k6 run` or `artillery run` |
+| Security | package.json | `npm audit` |
+
+**Mobile fallbacks (if no custom config):**
+
+| Platform | Unit | E2E |
+|----------|------|-----|
+| React Native | `npm test` (jest) | `maestro test` or `detox test` |
+| Android | `./gradlew test` | `./gradlew connectedAndroidTest` |
+| iOS | `xcodebuild test` | `maestro test` |
+| Flutter | `flutter test` | `flutter test integration_test/` |
+
+**IMPORTANT:** Always check `proagents.config.yaml` FIRST before using defaults.
+
 ### AI Platform Management
 | Command | Action |
 |---------|--------|