@kood/claude-code 0.1.0

package/templates/hono/docs/skills/gemini-review/references/prompt-templates.md

@@ -0,0 +1,303 @@
+# Prompt Templates for Gemini Review
+
+## Command Patterns
+
+All commands use the fixed model `gemini-3-pro-preview` with JSON output for parsing.
+
+### Basic Pattern
+```bash
+gemini -m gemini-3-pro-preview -p "{prompt}" --output-format json
+```
+
+### File Piping Pattern
+```bash
+cat {file_path} | gemini -m gemini-3-pro-preview -p "{instructions}" --output-format json
+```
+
+### Multi-line Heredoc Pattern
+```bash
+gemini -m gemini-3-pro-preview --output-format json -p "$(cat << 'EOF'
+{multi_line_prompt}
+EOF
+)"
+```
+
+### Response Parsing
+```bash
+result=$(gemini -m gemini-3-pro-preview -p "..." --output-format json)
+review_content=$(echo "$result" | jq -r '.response')
+echo "$review_content"
+```
+
+---
+
+## Plan Review Template
+
+```bash
+gemini -m gemini-3-pro-preview --output-format json -p "$(cat << 'EOF'
+[PLAN REVIEW REQUEST]
+
+## Context
+Project: {project_name}
+Tech Stack: {tech_stack}
+Current State: {current_state_description}
+
+## Implementation Plan
+{plan_content}
+
+## Review Checklist
+{checklist_items}
+
+## Review Instructions
+Analyze this implementation plan and provide feedback on:
+
+1. **Completeness**
+   - Are all requirements addressed?
+   - Any missing steps or considerations?
+
+2. **Logic & Feasibility**
+   - Will this approach work as intended?
+   - Any logical flaws or contradictions?
+
+3. **Edge Cases**
+   - What edge cases might be missed?
+   - How should they be handled?
+
+4. **Risks & Issues**
+   - Potential problems during implementation?
+   - Dependencies or blockers?
+
+5. **Alternatives**
+   - Better approaches available?
+   - Trade-offs to consider?
+
+Respond with:
+- ✅ Strengths of the plan
+- ⚠️ Concerns or issues found
+- 💡 Suggestions for improvement
+- 🔄 Alternative approaches (if any)
+EOF
+)"
+```
+
+---
+
+## Code Review Template
+
+### Option A: Inline Code
+```bash
+gemini -m gemini-3-pro-preview --output-format json -p "$(cat << 'EOF'
+[CODE REVIEW REQUEST]
+
+## Context
+File: {file_path}
+Language: {language}
+Tech Stack: {tech_stack}
+Purpose: {purpose_description}
+
+## Code to Review
+```{language}
+{code_content}
+```
+
+## Review Checklist
+{checklist_items}
+
+## Review Instructions
+Perform a thorough code review focusing on:
+
+1. **Bugs & Logic Errors**
+   - Incorrect logic or algorithms
+   - Off-by-one errors
+   - Null/undefined handling
+   - Type mismatches
+
+2. **Security Vulnerabilities**
+   - Injection risks (SQL, XSS, etc.)
+   - Authentication/authorization gaps
+   - Data exposure risks
+   - Input validation issues
+
+3. **Performance Issues**
+   - Inefficient algorithms (specify Big-O if problematic)
+   - N+1 queries
+   - Memory leaks
+   - Unnecessary operations
+
+4. **Best Practices**
+   - Code organization
+   - Naming conventions
+   - Error handling patterns
+   - Testing considerations
+
+5. **Maintainability**
+   - Code clarity
+   - Documentation needs
+   - Coupling and cohesion
+   - Future extensibility
+
+For each issue found, provide:
+- 🔴 CRITICAL / 🟡 IMPORTANT / 🟢 MINOR
+- Location (line number or function name)
+- Description of the issue
+- Suggested fix with code example
+EOF
+)"
+```
+
+### Option B: Piping File Content
+```bash
+cat {file_path} | gemini -m gemini-3-pro-preview --output-format json -p "Review this {language} code ({tech_stack} project) for:
+1. Bugs and logic errors
+2. Security vulnerabilities
+3. Performance issues
+4. Best practices violations
+5. Maintainability concerns
+
+For each issue: specify severity (🔴 CRITICAL/🟡 IMPORTANT/🟢 MINOR), location, description, and suggested fix with code."
+```
+
+---
+
+## Architecture Review Template
+
+```bash
+gemini -m gemini-3-pro-preview --output-format json -p "$(cat << 'EOF'
+[ARCHITECTURE REVIEW REQUEST]
+
+## System Overview
+Name: {system_name}
+Purpose: {system_purpose}
+Tech Stack: {tech_stack}
+Scale: {expected_scale}
+
+## Current/Proposed Architecture
+{architecture_description}
+
+## Components
+{components_list}
+
+## Data Flow
+{data_flow_description}
+
+## Specific Questions
+{specific_questions}
+
+## Review Instructions
+Evaluate this architecture considering:
+
+1. **Scalability**
+   - Can it handle expected load?
+   - Horizontal vs vertical scaling options?
+   - Bottlenecks identified?
+
+2. **Reliability**
+   - Single points of failure?
+   - Fault tolerance mechanisms?
+   - Recovery strategies?
+
+3. **Maintainability**
+   - Component coupling?
+   - Deployment complexity?
+   - Operational overhead?
+
+4. **Security**
+   - Attack surface?
+   - Data protection?
+   - Access control boundaries?
+
+5. **Cost Efficiency**
+   - Resource utilization?
+   - Scaling costs?
+   - Operational costs?
+
+6. **Trade-offs**
+   - What are we sacrificing?
+   - Alternative architectures?
+   - Migration path considerations?
+
+Respond with:
+- 📊 Overall assessment
+- ✅ Architectural strengths
+- ⚠️ Concerns and risks
+- 💡 Recommendations
+- 🔄 Alternative approaches worth considering
+EOF
+)"
+```
+
+---
+
+## Hono-Specific Additions
+
+When tech_stack is `hono`, append to the relevant template:
+
+```
+## Hono-Specific Review Points
+- Middleware chain correctness and order
+- zValidator usage and Zod schema design
+- HTTPException usage for error handling
+- Context (c) usage patterns
+- Type-safe bindings and variables
+- RPC client type exports
+- Built-in middleware configuration
+```
+
+---
+
+## Cloudflare Workers-Specific Additions
+
+When tech_stack is `cloudflare`, append to the relevant template:
+
+```
+## Cloudflare Workers-Specific Review Points
+- Edge runtime compatibility (no Node.js APIs)
+- KV, D1, R2 binding usage
+- wrangler.toml configuration
+- Environment variable handling via c.env
+- Request/response size limits
+- CPU and memory constraints
+- Subrequest limits (<50)
+- Durable Objects if applicable
+```
+
+---
+
+## Quick Review Templates
+
+### Security Quick Check
+```bash
+cat {file_path} | gemini -m gemini-3-pro-preview -p "Security audit this Hono code. Find: injection vulnerabilities, auth issues, data exposure risks, input validation gaps. List each with severity and fix." --output-format json
+```
+
+### Performance Quick Check
+```bash
+cat {file_path} | gemini -m gemini-3-pro-preview -p "Performance review this Hono code. Find: inefficient algorithms (note Big-O), N+1 queries, memory issues, unnecessary operations. Suggest optimizations." --output-format json
+```
+
+### Bug Hunt
+```bash
+cat {file_path} | gemini -m gemini-3-pro-preview -p "Find bugs in this code: logic errors, edge cases, null handling, race conditions, type issues. For each bug, show the fix." --output-format json
+```
+
+### Middleware Review
+```bash
+cat {file_path} | gemini -m gemini-3-pro-preview -p "Review this Hono middleware for: proper next() handling, error propagation, type safety, reusability, and side effects. Suggest improvements." --output-format json
+```
+
+### Validation Review
+```bash
+cat {file_path} | gemini -m gemini-3-pro-preview -p "Review Zod schemas and zValidator usage: schema completeness, error messages, type inference, reusability. Use Zod v4 syntax." --output-format json
+```
+
+---
+
+## Template Usage Instructions
+
+1. Select the appropriate base template (plan/code/architecture)
+2. Fill in all `{placeholder}` fields with actual content
+3. Load the relevant checklist from `checklists.md` based on tech stack
+4. Append tech-stack-specific additions if applicable
+5. Execute via Gemini CLI
+6. Parse JSON response to extract `.response` field
+7. Present both raw Gemini response and Claude's analysis

package/templates/tanstack-start/CLAUDE.md

@@ -0,0 +1,279 @@
+# CLAUDE.md - Claude Code Instructions
+
+> TanStack Start + React 프로젝트 작업 지침
+
+---
+
+## 🚨 STOP - 작업 전 필수 확인
+
+```
+┌─────────────────────────────────────────────────────────────┐
+│  이 프로젝트에서 작업하기 전에 이 문서를 끝까지 읽으세요.  │
+│  특히 ⛔ NEVER DO 섹션의 규칙은 절대 위반하지 마세요.      │
+│                                                             │
+│  📖 작업 유형별 상세 문서: docs/ 폴더 참조                  │
+└─────────────────────────────────────────────────────────────┘
+```
+
+---
+
+## ⛔ NEVER DO (절대 금지 - 예외 없음)
+
+### Git 커밋 금지 사항
+```
+❌ "Generated with Claude Code" 포함 금지
+❌ "🤖" 또는 AI 관련 이모지 포함 금지
+❌ "Co-Authored-By:" 헤더 포함 금지
+❌ AI/봇이 작성했다는 어떤 표시도 금지
+❌ 커밋 메시지 여러 줄 작성 금지
+❌ 커밋 메시지에 이모지 사용 금지
+```
+
+### Prisma 금지 사항
+```
+❌ prisma db push 자동 실행 금지
+❌ prisma migrate 자동 실행 금지
+❌ prisma generate 자동 실행 금지
+❌ schema.prisma 임의 변경 금지 (요청된 것만)
+```
+
+### API 구현 금지 사항
+```
+❌ /api 라우터에 함수 생성 금지 (사용자 명시 요청 제외)
+❌ Server Function에 inputValidator 누락 금지 (POST/PUT/PATCH)
+❌ Server Function에 middleware 누락 금지 (인증 필요 시)
+❌ handler 내부에서 수동 검증 금지 (inputValidator 사용)
+❌ handler 내부에서 수동 인증 체크 금지 (middleware 사용)
+```
+
+### 코드 검색 금지 사항
+```
+❌ grep, rg 등 기본 검색 도구 사용 금지
+❌ find 명령어로 코드 검색 금지
+✅ 코드베이스 검색 시 sgrep 사용 필수
+```
+
+---
+
+## ✅ ALWAYS DO (필수 실행)
+
+### 1. 작업 전: 관련 문서 읽기
+```
+UI 작업      → docs/design/ 읽기
+API 작업     → docs/library/tanstack-start/ 읽기
+DB 작업      → docs/library/prisma/ 읽기
+인증 작업    → docs/library/better-auth/ 읽기
+```
+
+### 2. MCP 도구 적극 활용
+```
+코드베이스 검색     → sgrep 사용 (grep/rg 금지)
+복잡한 분석/디버깅  → Sequential Thinking 사용
+라이브러리 문서     → Context7 사용
+```
+**상세**: `docs/mcp/` 참고
+
+### 3. 복잡한 작업 시: Gemini Review 실행
+```
+아키텍처 설계/변경  → gemini-review (architecture)
+구현 계획 검증      → gemini-review (plan)
+복잡한 코드 리뷰    → gemini-review (code)
+```
+
+**실행 조건**:
+- 3개 이상 파일 수정하는 기능 구현
+- 새로운 아키텍처 패턴 도입
+- 보안 관련 코드 (인증, 권한, 암호화)
+- 성능 크리티컬 코드
+
+**상세**: `docs/skills/gemini-review/SKILL.md` 참고
+
+### 4. 작업 완료 후: Git 커밋
+```bash
+git add .
+git commit -m "<prefix>: <설명>"
+```
+
+**커밋 형식**: `<prefix>: <설명>` (한 줄, 본문 없음)
+
+**Prefix**: `feat` | `fix` | `refactor` | `style` | `docs` | `test` | `chore` | `perf` | `ci`
+
+**예시**:
+```bash
+feat: 사용자 로그인 기능 추가
+fix: 세션 만료 오류 수정
+docs: API 문서 업데이트
+```
+
+---
+
+## 📚 문서 참조 테이블
+
+| 작업 | 문서 경로 | 필독 여부 |
+|------|----------|----------|
+| **전체 가이드** | `docs/README.md` | 🔴 필수 |
+| **Git 규칙** | `docs/git/index.md` | 🔴 필수 |
+| **MCP 도구** | `docs/mcp/` | 🔴 필수 |
+| **Gemini Review** | `docs/skills/gemini-review/` | 🟡 복잡한 작업 시 |
+| **UI 개발** | `docs/design/` | 🟡 해당 시 |
+| **API 개발** | `docs/library/tanstack-start/` | 🟡 해당 시 |
+| **인증** | `docs/library/better-auth/` | 🟡 해당 시 |
+| **DB** | `docs/library/prisma/` | 🟡 해당 시 |
+| **데이터 페칭** | `docs/library/tanstack-query/` | 🟡 해당 시 |
+| **검증** | `docs/library/zod/` | 🟡 해당 시 |
+| **배포** | `docs/deployment/` | 🟡 해당 시 |
+
+---
+
+## 🛠 Tech Stack (버전 주의)
+
+| 기술 | 버전 | 주의사항 |
+|------|------|----------|
+| TanStack Start | 최신 | Framework |
+| TypeScript | 5.x | strict mode |
+| Tailwind CSS | 4.x | `@theme` 사용 |
+| Prisma | **7.x** | `prisma-client` (js 아님), output 필수 |
+| Zod | **4.x** | `z.email()`, `z.url()` (string().email() 아님) |
+| Better Auth | 최신 | 인증 |
+| TanStack Query | 5.x | 데이터 페칭 |
+
+---
+
+## 📁 Directory Structure
+
+```
+src/
+├── routes/                   # File-based routes
+│   ├── __root.tsx            # Root layout
+│   ├── index.tsx             # / (Home)
+│   ├── $slug.tsx             # Dynamic route
+│   └── users/
+│       ├── index.tsx         # /users
+│       ├── -components/      # 페이지 전용 컴포넌트
+│       ├── -hooks/           # 페이지 전용 훅
+│       └── -functions/       # 페이지 전용 Server Functions ⭐
+├── functions/                # 공통 Server Functions ⭐
+│   ├── auth.ts               # 인증 관련
+│   └── user.ts               # 사용자 관련
+├── components/ui/            # 공통 UI 컴포넌트
+├── middleware/               # 공통 미들웨어
+│   └── auth.ts               # Better Auth 미들웨어
+├── database/prisma.ts        # Prisma Client
+└── lib/                      # 유틸리티
+```
+
+### Server Functions 위치 규칙 ⭐
+```
+공통 함수 (여러 라우트에서 사용)  → @/functions/
+라우트 전용 함수 (해당 라우트만)  → routes/[경로]/-functions/
+```
+
+**`-` prefix**: 라우트에서 제외되는 폴더
+
+---
+
+## 🔧 Code Conventions
+
+### File Naming
+- **kebab-case**: `user-profile.tsx`, `auth-service.ts`
+- **Routes**: `__root.tsx`, `$param.tsx`
+
+### TypeScript
+- `const` 선언 사용 (function 대신)
+- 명시적 return type
+- `interface` (객체) / `type` (유니온)
+- `any` 금지 → `unknown` 사용
+
+### Import
+```typescript
+// @/* → ./src/*
+import { prisma } from '@/database/prisma'
+import { getUsers } from '@/services/user'
+```
+
+**순서**: 외부 → 내부(@/) → 상대경로 → type imports
+
+---
+
+## 📝 Quick Patterns (복사용)
+
+### Server Function (GET + 인증)
+```typescript
+// ✅ 올바른 패턴: middleware 사용
+export const getUsers = createServerFn({ method: 'GET' })
+  .middleware([authMiddleware])  // ⭐ 인증 미들웨어 필수
+  .handler(async () => prisma.user.findMany())
+```
+
+### Server Function (POST + Validation + 인증)
+```typescript
+// ✅ 올바른 패턴: inputValidator + middleware 모두 사용
+export const createUser = createServerFn({ method: 'POST' })
+  .middleware([authMiddleware])        // ⭐ 인증 미들웨어 필수
+  .inputValidator(createUserSchema)    // ⭐ Zod 스키마 필수
+  .handler(async ({ data }) => prisma.user.create({ data }))
+```
+
+### ❌ 잘못된 패턴 (금지)
+```typescript
+// ❌ handler 내부에서 수동 검증 금지
+export const createUser = createServerFn({ method: 'POST' })
+  .handler(async ({ data }) => {
+    // ❌ 이렇게 하지 마세요!
+    if (!data.email) throw new Error('Email required')
+    const session = await getSession()  // ❌ 수동 인증 체크 금지
+    // ...
+  })
+```
+
+### Zod Schema (v4 문법!)
+```typescript
+const schema = z.object({
+  email: z.email(),           // ✅ v4
+  name: z.string().min(1).trim(),
+  website: z.url().optional(), // ✅ v4
+})
+```
+
+### Route with Loader
+```tsx
+export const Route = createFileRoute('/users')({
+  component: UsersPage,
+  loader: async () => ({ users: await getUsers() }),
+})
+```
+
+### TanStack Query
+```tsx
+const { data } = useQuery({
+  queryKey: ['users'],
+  queryFn: () => getUsers(),
+})
+
+const mutation = useMutation({
+  mutationFn: createUser,
+  onSuccess: () => queryClient.invalidateQueries({ queryKey: ['users'] }),
+})
+```
+
+---
+
+## 🎨 UI/UX Rules
+
+- **색상**: 60-30-10 규칙 (배경-보조-강조)
+- **간격**: 8px 그리드 (8의 배수)
+- **접근성**: WCAG AA (대비 4.5:1+)
+- **폰트**: 최대 2-3개
+- **Safe Area**: `tailwindcss-safe-area` 사용
+
+**상세**: `docs/design/` 참고
+
+---
+
+## 🔗 Quick Links
+
+- [문서 가이드](./docs/README.md)
+- [Git 규칙](./docs/git/index.md)
+- [MCP 가이드](./docs/mcp/index.md)
+- [디자인 가이드](./docs/design/index.md)
+- [아키텍처](./docs/architecture/architecture.md)