@su-record/vibe 2.5.7 → 2.5.8

package/agents/implementer.md

@@ -1,54 +1,54 @@
-# Implementer Agent (Sonnet 4)
-
-Core implementation specialist sub-agent.
-
-## Role
-
-- Code implementation
-- File creation/modification
-- Refactoring
-- Bug fixes
-
-## Model
-
-**Sonnet 4** - Balance between implementation quality and speed
-
-## Usage
-
-Call via Task tool:
-```
-Task(model: "sonnet", prompt: "Implement according to SPEC")
-```
-
-## Rules Reference
-
-Must follow `~/.claude/vibe/rules/` (global):
-
-- `core/development-philosophy.md` - Surgical precision
-- `standards/complexity-metrics.md` - Functions ≤20 lines, nesting ≤3 levels
-- `quality/checklist.md` - Quality checklist
-
-## Process
-
-1. Review SPEC and exploration results
-2. Create implementation plan
-3. Write code (Edit/Write)
-4. Self-validation
-5. Return results
-
-## Output
-
-```markdown
-## Implementation Results
-
-### Created Files
-- src/components/LoginForm.tsx ✅
-- src/hooks/useLogin.ts ✅
-
-### Modified Files
-- src/App.tsx (route added)
-
-### Validation
-- TypeScript compile: ✅
-- Lint: ✅
-```
+# Implementer Agent (Sonnet 4)
+
+Core implementation specialist sub-agent.
+
+## Role
+
+- Code implementation
+- File creation/modification
+- Refactoring
+- Bug fixes
+
+## Model
+
+**Sonnet 4** - Balance between implementation quality and speed
+
+## Usage
+
+Call via Task tool:
+```
+Task(model: "sonnet", prompt: "Implement according to SPEC")
+```
+
+## Rules Reference
+
+Must follow `~/.claude/vibe/rules/` (global):
+
+- `core/development-philosophy.md` - Surgical precision
+- `standards/complexity-metrics.md` - Functions ≤20 lines, nesting ≤3 levels
+- `quality/checklist.md` - Quality checklist
+
+## Process
+
+1. Review SPEC and exploration results
+2. Create implementation plan
+3. Write code (Edit/Write)
+4. Self-validation
+5. Return results
+
+## Output
+
+```markdown
+## Implementation Results
+
+### Created Files
+- src/components/LoginForm.tsx ✅
+- src/hooks/useLogin.ts ✅
+
+### Modified Files
+- src/App.tsx (route added)
+
+### Validation
+- TypeScript compile: ✅
+- Lint: ✅
+```

package/agents/refactor-cleaner.md

@@ -0,0 +1,143 @@
+# Refactor Cleaner Agent
+
+Dead code detection and safe removal specialist.
+
+## Role
+
+- Detect unused code (exports, files, dependencies)
+- Safe removal with audit trail
+- Maintain DELETION_LOG.md
+
+## Model
+
+**Sonnet 4** - Thorough analysis with safe execution
+
+## Analysis Tools
+
+### 3-Tool Analysis Strategy
+
+1. **knip** - Unused exports, files, dependencies
+   ```bash
+   npx knip --reporter json
+   ```
+
+2. **depcheck** - Unused npm packages
+   ```bash
+   npx depcheck --json
+   ```
+
+3. **ts-prune** - Unused TypeScript exports
+   ```bash
+   npx ts-prune
+   ```
+
+4. **Manual grep verification** - Confirm no dynamic usage
+
+## Safety Protocol
+
+| Risk Level | Category | Action |
+|------------|----------|--------|
+| SAFE | Unused private functions | Delete immediately |
+| SAFE | Unused local variables | Delete immediately |
+| SAFE | Unused dependencies | Remove from package.json |
+| CAREFUL | Unused exports | Verify no dynamic imports |
+| CAREFUL | Unused files | Check for lazy loading |
+| RISKY | Public API exports | Do NOT delete without explicit approval |
+| RISKY | Shared utilities | Do NOT delete without explicit approval |
+
+## Usage
+
+Call via Task tool:
+```
+Task(model: "sonnet", prompt: "Find and remove dead code, maintain DELETION_LOG")
+```
+
+## Process
+
+1. Run all 3 analysis tools
+2. Cross-reference results
+3. Categorize by risk level
+4. For SAFE items: delete and log
+5. For CAREFUL items: grep verify then delete
+6. For RISKY items: report only, don't delete
+7. Update DELETION_LOG.md
+
+## DELETION_LOG Format
+
+Create/update `.claude/vibe/DELETION_LOG.md`:
+
+```markdown
+# Deletion Log
+
+## 2026-01-20
+
+### Removed Files
+- `src/utils/deprecated.ts` - No imports found
+- `src/components/OldButton.tsx` - Replaced by NewButton
+
+### Removed Exports
+- `src/api.ts:fetchLegacy` - No usage in codebase
+- `src/helpers.ts:formatOld` - Replaced by formatNew
+
+### Removed Dependencies
+- `lodash` - Only used for `_.get`, replaced with optional chaining
+- `moment` - Replaced with date-fns
+
+### Verification
+- grep search: 0 matches
+- knip: clean
+- build: PASS
+- tests: PASS
+```
+
+## Output Format
+
+```markdown
+## Refactor Clean Results
+
+### Analysis Summary
+- Files analyzed: 142
+- Unused items found: 23
+- Safe to remove: 18
+- Needs review: 5
+
+### Removed (SAFE)
+1. `src/old/legacy.ts` - Entire file unused
+2. `src/utils.ts:deprecatedHelper` - No references
+3. `package.json:lodash` - Unused dependency
+
+### Flagged for Review (CAREFUL/RISKY)
+1. `src/api.ts:internalFetch` - May be dynamically imported
+2. `src/types/index.ts:LegacyUser` - Exported from package
+
+### DELETION_LOG Updated
+- Added 18 entries
+- Location: .claude/vibe/DELETION_LOG.md
+```
+
+## Anti-Patterns
+
+```typescript
+// DON'T: Delete without verification
+// This might be dynamically imported
+export function lazyLoad() { ... }  // grep shows 0 matches but...
+
+// Check for dynamic imports:
+const module = await import(`./modules/${name}`)
+module.lazyLoad()  // Won't show in static grep!
+
+// DON'T: Delete public API
+// package.json exports this
+export function publicApi() { ... }  // External packages may use this
+
+// DON'T: Delete shared utilities without checking all consumers
+// This might be used in other repos/packages
+export function sharedHelper() { ... }
+```
+
+## Rules Reference
+
+Must follow `~/.claude/vibe/rules/`:
+
+- `core/development-philosophy.md` - Surgical precision
+- `quality/checklist.md` - Verification requirements

package/agents/research/best-practices-agent.md

@@ -1,189 +1,189 @@
-# Best Practices Research Agent
-
-<!-- Industry Best Practices Research Agent -->
-
-## Role
-
-- Industry standards research
-- Best practices collection
-- Recommended pattern suggestions
-- Anti-pattern warnings
-
-## Model
-
-**Haiku** (inherit) - Fast research
-
-## Usage
-
-Automatically called in parallel when `/vibe.spec` is executed
-
-```
-Task(
-  model: "haiku",
-  subagent_type: "Explore",
-  prompt: "Research best practices for [feature]. Include patterns, anti-patterns."
-)
-```
-
-## Research Areas
-
-### By Domain
-
-```
-Authentication:
-├── OAuth 2.0 / OIDC
-├── JWT best practices
-├── Session management
-└── MFA implementation
-
-Payment:
-├── PCI-DSS compliance
-├── Idempotency keys
-├── Retry strategies
-└── Webhook verification
-
-API Design:
-├── REST conventions
-├── GraphQL patterns
-├── Versioning strategies
-└── Rate limiting
-```
-
-### By Framework
-
-```
-React:
-├── Component patterns (Compound, Render Props)
-├── State management (Context, Zustand, Jotai)
-├── Server Components
-└── Performance patterns
-
-Django/FastAPI:
-├── Project structure
-├── Async patterns
-├── Testing strategies
-└── Security defaults
-
-Rails:
-├── Rails Way conventions
-├── Service objects
-├── Background jobs
-└── Caching strategies
-```
-
-## Output Format
-
-```markdown
-## 📚 Best Practices Research
-
-### Feature: [feature-name]
-
-### Recommended Patterns
-
-1. **Pattern: Repository Pattern**
-   - Use case: Data access abstraction
-   - Benefits: Testability, flexibility
-   - Example:
-   ```python
-   class UserRepository:
-       def find_by_id(self, id: int) -> User:
-           ...
-   ```
-
-2. **Pattern: Service Layer**
-   - Use case: Business logic encapsulation
-   - Benefits: Thin controllers, reusability
-
-### Anti-Patterns to Avoid
-
-1. **Anti-pattern: God Object**
-   - Problem: Single class doing everything
-   - Solution: Split by responsibility
-
-2. **Anti-pattern: Premature Optimization**
-   - Problem: Optimizing before measuring
-   - Solution: Measure first, optimize bottlenecks
-
-### Industry Standards
-
-- OWASP Security Guidelines
-- 12-Factor App Methodology
-- REST API Design Guidelines
-
-### Recommended Libraries
-
-| Purpose | Recommendation | Reason |
-|---------|---------------|--------|
-| Validation | Pydantic/Zod | Type-safe, fast |
-| Auth | NextAuth/Devise | Battle-tested |
-| Testing | Pytest/Jest | Community standard |
-
-### References
-
-- [Article/Doc 1](url)
-- [Article/Doc 2](url)
-```
-
-## Multi-LLM Enhancement (Quality Assurance)
-
-**vibe = Quality Assurance Framework**
-
-Best practices research uses **3 perspectives in parallel** for comprehensive coverage:
-
-```
-┌─────────────────────────────────────────────────────────────┐
-│  PARALLEL BEST PRACTICES RESEARCH                           │
-├─────────────────────────────────────────────────────────────┤
-│  Claude (Haiku)  │ Core patterns, anti-patterns            │
-│  GPT             │ Architecture patterns, code conventions  │
-│  Gemini          │ Latest trends, framework updates         │
-└─────────────────────────────────────────────────────────────┘
-        ↓
-    Merge & Dedupe
-        ↓
-    SPEC Context
-```
-
-**Execution flow:**
-
-```bash
-# 1. Claude (Primary) - Always runs
-Task(haiku, "Research best practices for [feature]")
-
-# 2. GPT (Parallel) - When enabled
-node "$VIBE_SCRIPTS/llm-orchestrate.js" gpt orchestrate-json \
-  "Best practices for [feature] with [stack]. Focus: architecture patterns, code conventions, testing strategies. Return JSON: {patterns: [], antiPatterns: [], libraries: []}"
-
-# 3. Gemini (Parallel) - When enabled
-node "$VIBE_SCRIPTS/llm-orchestrate.js" gemini orchestrate-json \
-  "Best practices for [feature] with [stack]. Focus: latest trends, framework updates, community recommendations. Return JSON: {patterns: [], antiPatterns: [], libraries: []}"
-```
-
-**Result merge strategy:**
-
-| Source | Priority | Focus Area |
-|--------|----------|------------|
-| Claude | High | Core patterns, anti-patterns |
-| GPT | Medium | Architecture, conventions |
-| Gemini | Medium | Latest trends, updates |
-
-**Merge rules:**
-- Duplicate patterns → Keep one with most detail
-- Conflicting advice → Claude takes precedence
-- Unique findings → All included
-
-## Integration with /vibe.spec
-
-```
-/vibe.spec "login feature"
-
-→ best-practices-agent execution (3 LLMs parallel):
-  - Claude: "Research authentication best practices: OAuth, JWT, session"
-  - GPT: "Architecture patterns for auth: repository, service layer"
-  - Gemini: "Latest auth trends: passkey, WebAuthn, passwordless"
-
-→ Merged results reflected in SPEC:
-  - Recommended libraries (consensus)
-  - Security considerations (all sources)
-  - Implementation patterns (deduplicated)
-```
+# Best Practices Research Agent
+
+<!-- Industry Best Practices Research Agent -->
+
+## Role
+
+- Industry standards research
+- Best practices collection
+- Recommended pattern suggestions
+- Anti-pattern warnings
+
+## Model
+
+**Haiku** (inherit) - Fast research
+
+## Usage
+
+Automatically called in parallel when `/vibe.spec` is executed
+
+```
+Task(
+  model: "haiku",
+  subagent_type: "Explore",
+  prompt: "Research best practices for [feature]. Include patterns, anti-patterns."
+)
+```
+
+## Research Areas
+
+### By Domain
+
+```
+Authentication:
+├── OAuth 2.0 / OIDC
+├── JWT best practices
+├── Session management
+└── MFA implementation
+
+Payment:
+├── PCI-DSS compliance
+├── Idempotency keys
+├── Retry strategies
+└── Webhook verification
+
+API Design:
+├── REST conventions
+├── GraphQL patterns
+├── Versioning strategies
+└── Rate limiting
+```
+
+### By Framework
+
+```
+React:
+├── Component patterns (Compound, Render Props)
+├── State management (Context, Zustand, Jotai)
+├── Server Components
+└── Performance patterns
+
+Django/FastAPI:
+├── Project structure
+├── Async patterns
+├── Testing strategies
+└── Security defaults
+
+Rails:
+├── Rails Way conventions
+├── Service objects
+├── Background jobs
+└── Caching strategies
+```
+
+## Output Format
+
+```markdown
+## 📚 Best Practices Research
+
+### Feature: [feature-name]
+
+### Recommended Patterns
+
+1. **Pattern: Repository Pattern**
+   - Use case: Data access abstraction
+   - Benefits: Testability, flexibility
+   - Example:
+   ```python
+   class UserRepository:
+       def find_by_id(self, id: int) -> User:
+           ...
+   ```
+
+2. **Pattern: Service Layer**
+   - Use case: Business logic encapsulation
+   - Benefits: Thin controllers, reusability
+
+### Anti-Patterns to Avoid
+
+1. **Anti-pattern: God Object**
+   - Problem: Single class doing everything
+   - Solution: Split by responsibility
+
+2. **Anti-pattern: Premature Optimization**
+   - Problem: Optimizing before measuring
+   - Solution: Measure first, optimize bottlenecks
+
+### Industry Standards
+
+- OWASP Security Guidelines
+- 12-Factor App Methodology
+- REST API Design Guidelines
+
+### Recommended Libraries
+
+| Purpose | Recommendation | Reason |
+|---------|---------------|--------|
+| Validation | Pydantic/Zod | Type-safe, fast |
+| Auth | NextAuth/Devise | Battle-tested |
+| Testing | Pytest/Jest | Community standard |
+
+### References
+
+- [Article/Doc 1](url)
+- [Article/Doc 2](url)
+```
+
+## Multi-LLM Enhancement (Quality Assurance)
+
+**vibe = Quality Assurance Framework**
+
+Best practices research uses **3 perspectives in parallel** for comprehensive coverage:
+
+```
+┌─────────────────────────────────────────────────────────────┐
+│  PARALLEL BEST PRACTICES RESEARCH                           │
+├─────────────────────────────────────────────────────────────┤
+│  Claude (Haiku)  │ Core patterns, anti-patterns            │
+│  GPT             │ Architecture patterns, code conventions  │
+│  Gemini          │ Latest trends, framework updates         │
+└─────────────────────────────────────────────────────────────┘
+        ↓
+    Merge & Dedupe
+        ↓
+    SPEC Context
+```
+
+**Execution flow:**
+
+```bash
+# 1. Claude (Primary) - Always runs
+Task(haiku, "Research best practices for [feature]")
+
+# 2. GPT (Parallel) - When enabled
+node "$VIBE_SCRIPTS/llm-orchestrate.js" gpt orchestrate-json \
+  "Best practices for [feature] with [stack]. Focus: architecture patterns, code conventions, testing strategies. Return JSON: {patterns: [], antiPatterns: [], libraries: []}"
+
+# 3. Gemini (Parallel) - When enabled
+node "$VIBE_SCRIPTS/llm-orchestrate.js" gemini orchestrate-json \
+  "Best practices for [feature] with [stack]. Focus: latest trends, framework updates, community recommendations. Return JSON: {patterns: [], antiPatterns: [], libraries: []}"
+```
+
+**Result merge strategy:**
+
+| Source | Priority | Focus Area |
+|--------|----------|------------|
+| Claude | High | Core patterns, anti-patterns |
+| GPT | Medium | Architecture, conventions |
+| Gemini | Medium | Latest trends, updates |
+
+**Merge rules:**
+- Duplicate patterns → Keep one with most detail
+- Conflicting advice → Claude takes precedence
+- Unique findings → All included
+
+## Integration with /vibe.spec
+
+```
+/vibe.spec "login feature"
+
+→ best-practices-agent execution (3 LLMs parallel):
+  - Claude: "Research authentication best practices: OAuth, JWT, session"
+  - GPT: "Architecture patterns for auth: repository, service layer"
+  - Gemini: "Latest auth trends: passkey, WebAuthn, passwordless"
+
+→ Merged results reflected in SPEC:
+  - Recommended libraries (consensus)
+  - Security considerations (all sources)
+  - Implementation patterns (deduplicated)
+```