Skip to content

claude-config-reference.md

Latest commit

 

History

History
244 lines (185 loc) · 8.56 KB

claude-config-reference.md

File metadata and controls

244 lines (185 loc) · 8.56 KB

claude-config Reference Guide

This document explains how AD-SDLC agents can reference resources from claude-config to improve code quality.

Overview

claude-config is a Claude Code configuration management and development guidelines system. When AD-SDLC agents reference these guidelines during code generation, review, and documentation, they can produce consistent, high-quality results.

Key Resources

Skills (Auto-loading Guidelines)

Skill Description Path
coding-guidelines Coding standards, quality, error handling .claude/skills/coding-guidelines/
security-audit Security guidelines, OWASP Top 10 .claude/skills/security-audit/
performance-review Performance optimization, profiling .claude/skills/performance-review/
api-design API design, architecture .claude/skills/api-design/
project-workflow Git, issues, PR management .claude/skills/project-workflow/
documentation README, API docs, comments .claude/skills/documentation/

Rules (.claude/rules/)

Rules are loaded automatically by Claude Code based on alwaysApply and paths frontmatter.

.claude/rules/
├── core/
│   ├── principles.md       # Think, Minimize, Verify (alwaysApply)
│   ├── communication.md    # English code, Korean conversation (alwaysApply)
│   └── environment.md      # KST, Korean locale (alwaysApply)
├── coding/
│   ├── standards.md        # Naming, modularity, SOLID, quality
│   ├── error-handling.md   # Error patterns, validation
│   ├── performance.md      # Profiling, caching, algorithms
│   ├── safety.md           # Safe code practices
│   ├── implementation-standards.md  # Tier implementation
│   └── cpp-specifics.md    # C++ specific rules
├── api/
│   ├── api-design.md       # REST, GraphQL design
│   ├── architecture.md     # SOLID, layered design
│   ├── observability.md    # Logging, metrics, tracing
│   └── rest-api.md         # REST conventions
├── project-management/
│   ├── build.md            # Build, dependency management
│   ├── testing.md          # Unit/integration/E2E tests
│   └── documentation.md    # API docs, README
├── security.md             # Input validation, auth, OWASP Top 10
└── workflow/
    ├── git-commit-format.md    # Conventional Commits
    ├── github-issue-5w1h.md    # Issue framework
    └── github-pr-5w1h.md       # PR framework

Note: coding-standards/ was renamed to coding/, and api-architecture/ to api/. quality.md and concurrency.md were merged into standards.md. logging.md was merged into observability.md.

Last reconciled: 2026-04-11

Recommended References by Agent

Worker Agent

Responsible for code implementation. Requires the most guideline references.

Reference Resource Purpose
coding/standards.md Naming rules, code structure, quality
coding/error-handling.md Exception handling patterns
project-management/testing.md Test writing rules (AAA pattern)
coding/implementation-standards.md Complete tier implementation

Prompt Example:

## Implementation Guidelines

Follow these standards from claude-config:

### Naming Conventions
- Variables: camelCase (e.g., `userName`, `itemCount`)
- Classes: PascalCase (e.g., `UserService`, `OrderItem`)
- Constants: UPPER_SNAKE_CASE (e.g., `MAX_RETRIES`)
- Files: kebab-case (e.g., `user-service.ts`)

### Error Handling
- Use specific error types, not generic Error
- Include actionable error messages
- Distinguish recoverable vs non-recoverable errors

### Testing (AAA Pattern)
- Arrange: Set up test data and conditions
- Act: Execute the code under test
- Assert: Verify the expected outcome
- Target: 80%+ code coverage

PR Reviewer Agent

Responsible for code review. Focuses on quality and security guidelines.

Reference Resource Purpose
security.md Security vulnerability checklist
coding/standards.md Code quality criteria
coding/performance.md Performance issue identification

Prompt Example:

## Review Checklist

### Security (from claude-config/security.md)
- [ ] Input validation on all user inputs
- [ ] SQL injection prevention (parameterized queries)
- [ ] XSS prevention (output encoding)
- [ ] No hardcoded secrets or credentials
- [ ] Proper authentication/authorization checks

### Code Quality (from claude-config/coding/standards.md)
- [ ] Cyclomatic complexity < 10 per function
- [ ] No code duplication (DRY principle)
- [ ] Single Responsibility Principle
- [ ] Clear and meaningful names

SDS Writer Agent

Responsible for software design documentation. Focuses on API/architecture guidelines.

Reference Resource Purpose
api/api-design.md REST/GraphQL design patterns
api/architecture.md Architecture patterns, SOLID
api/observability.md Monitoring, health check design

Prompt Example:

## API Design Guidelines

### REST API (from claude-config)
- Resource-oriented URLs: `/users/{id}`, `/orders/{id}/items`
- Appropriate HTTP methods: GET (read), POST (create), PUT (update), DELETE (remove)
- Consistent response format with proper status codes
- Versioning strategy: URL path (v1) or header

### Architecture Patterns
- Layered architecture: Controller → Service → Repository
- Dependency Injection for testability
- Interface segregation for loose coupling

Controller Agent

Responsible for task distribution and issue management.

Reference Resource Purpose
project-management/build.md Build priority decisions
workflow/github-issue-5w1h.md Issue structuring

Document Writers (PRD/SRS/SDS)

Document writing agents.

Reference Resource Purpose
project-management/documentation.md Documentation standards
communication.md Language and style rules

Reference Methods

Method 1: GitHub Raw URL

# View SKILL.md
curl -s https://raw.githubusercontent.com/kcenon/claude-config/main/project/.claude/skills/coding-guidelines/SKILL.md

# View specific guideline
curl -s https://raw.githubusercontent.com/kcenon/claude-config/main/project/claude-config/rules/coding/general.md

Method 2: Local Clone

# Clone to same directory
git clone https://github.com/kcenon/claude-config.git ../claude-config

# Reference in agent prompt
# "Follow ../claude-config/project/.claude/rules/ when generating code"

Method 3: Direct Inclusion in Agent Prompts

Include core rules directly in agent definition files (.claude/agents/*.md):

---
name: worker
description: "Implements code based on assigned issues"
tools:
  - Read
  - Write
  - Edit
  - Bash
model: inherit
---

# Worker Agent

## Code Generation Standards

Follow these rules when generating code:

### From claude-config/coding
1. Naming: camelCase (variables), PascalCase (classes), UPPER_SNAKE_CASE (constants)
2. Error handling: Use specific error types, distinguish recoverability
3. Testing: AAA pattern, 80%+ coverage

### Reference Links
- [Full Coding Guidelines](https://github.com/kcenon/claude-config/blob/main/project/claude-config/rules/coding/general.md)
- [Error Handling](https://github.com/kcenon/claude-config/blob/main/project/claude-config/rules/coding/error-handling.md)

Expected Benefits

Aspect Benefit
Consistency All agents apply the same coding standards
Quality Code generation based on verified best practices
Security Proactive vulnerability prevention with systematic security review
Maintenance Guideline updates reflect across all agents

Cautions

  1. Selective Application: Choose only guidelines appropriate for your project
  2. Version Management: Verify compatibility when claude-config is updated
  3. Prompt Size: Including all guidelines increases token cost; include only essentials

Related Links