Skills System

# Analyzes last 30 commits:
Detected style: SEMANTIC (Korean)
- 22/30 commits use feat:/fix:/chore: prefix
- 18/30 commits in Korean

# Generated commits follow detected pattern:
feat: 사용자 모델 추가
test: 사용자 서비스 테스트 추가

/playwright Navigate to example.com and take a screenshot

// Skill loads MCP server
mcp:
  playwright:
    command: npx
    args: ["@playwright/mcp@latest"]

// Agent can now use playwright_* tools

task(
  category="visual-engineering",
  load_skills=["frontend-ui-ux"],
  prompt="Create a landing page hero section with bold typography and smooth scroll animations"
)

---
name: my-skill
description: "Custom skill for my project"
---

# My Custom Skill

This content will be injected into the agent's system prompt.

## Guidelines

- Specific instruction 1
- Specific instruction 2
- Pattern to follow from codebase

## Examples

\`\`\`typescript
// Example code showing the pattern
\`\`\`

## Anti-Patterns

- Never do X
- Avoid Y

---
name: api-client
description: "Internal API client with authentication"
mcp:
  internal-api:
    command: node
    args: ["./scripts/api-mcp-server.js"]
---

# Internal API Client Skill

You have access to our internal API via the `internal-api` MCP server.

## Available Operations

Use the `skill_mcp` tool to invoke API operations:

```typescript
skill_mcp(
  server_name="internal-api",
  operation="get_user",
  params={ user_id: 123 }
)

### Skill with OAuth MCP

**File:** `.opencode/skills/external-service/SKILL.md`

```markdown
---
name: external-service
description: "OAuth-protected external service"
mcp:
  external-api:
    url: https://api.example.com/mcp
    oauth:
      clientId: ${EXTERNAL_SERVICE_CLIENT_ID}
      scopes: ["read", "write"]
---

# External Service Skill

Access to external service API with OAuth 2.1 authentication.

## Pre-authentication

Authenticate before first use:

```bash
bunx oh-my-opencode mcp oauth login external-api \
  --server-url https://api.example.com

skill_mcp(
  server_name="external-api",
  operation="fetch_data",
  params={ query: "search term" }
)

**Environment variables:**

```bash
# .env
EXTERNAL_SERVICE_CLIENT_ID=your-client-id

task(
  category="deep",
  load_skills=["my-skill", "api-client"],
  prompt="Use the API client to fetch user data and process it"
)

skill(name="my-skill")
// Returns skill content with context applied

{
  "agents": {
    "sisyphus": {
      "skills": ["git-master", "my-skill"]
    }
  }
}

// Discover available operations
skill_mcp(server_name="playwright", operation="list_tools")

// Invoke operation
skill_mcp(
  server_name="playwright",
  operation="navigate",
  params={ url: "https://example.com" }
)

{
  "disabled_skills": ["playwright"]
}

{
  "skills": {
    "sources": [
      { 
        "path": "./custom-skills", 
        "recursive": true,
        "glob": "**/*.md"
      },
      "https://example.com/shared-skill.yaml"
    ]
  }
}

{
  "skills": {
    "my-skill": {
      "description": "Override description",
      "model": "openai/gpt-5.2",
      "agent": "custom-agent",
      "allowed-tools": ["read", "bash"],
      "subtask": true,
      "license": "MIT",
      "metadata": {
        "author": "Your Name",
        "version": "1.0.0"
      }
    }
  }
}

---
name: api-testing
description: "REST API testing with authentication"
---

# API Testing Skill

You are an API testing specialist.

## Testing Workflow

1. **Happy Path**: Test expected behavior first
2. **Edge Cases**: Test boundary conditions
3. **Error Cases**: Test all documented error responses
4. **Authentication**: Test with valid/invalid/expired tokens
5. **Rate Limiting**: Verify rate limit enforcement

## Test Structure

```typescript
describe('GET /api/users/:id', () => {
  describe('Happy Path', () => {
    it('returns user when authenticated', async () => {
      // Test implementation
    })
  })
  
  describe('Edge Cases', () => {
    it('handles user not found', async () => {
      // Test implementation
    })
  })
  
  describe('Authentication', () => {
    it('returns 401 when not authenticated', async () => {
      // Test implementation
    })
  })
})

**Usage:**
```typescript
task(
  category="test-writer",
  load_skills=["api-testing"],
  prompt="Write comprehensive API tests for /api/users endpoints"
)

---
name: db-migration
description: "Safe database migration practices"
---

# Database Migration Skill

## Safety Checklist

Before writing migration:

- [ ] Backup strategy defined
- [ ] Rollback tested
- [ ] No data loss possible
- [ ] Compatible with zero-downtime deployment
- [ ] Indexes created CONCURRENTLY (PostgreSQL)

## Migration Template

```sql
-- Up Migration
BEGIN;
  -- Changes here
  -- Use transactions
  -- Add constraints as NOT VALID first, validate separately
COMMIT;

-- Down Migration
BEGIN;
  -- Exact reverse
COMMIT;

-- Migration 1: Add column (nullable)
ALTER TABLE users ADD COLUMN email TEXT;

-- Migration 2: Backfill data (in batches)
UPDATE users SET email = old_email WHERE email IS NULL;

-- Migration 3: Add NOT NULL constraint
ALTER TABLE users ALTER COLUMN email SET NOT NULL;

-- Step 1: Deploy code ignoring column
-- Step 2: Wait 1 deployment cycle
-- Step 3: Drop column
ALTER TABLE users DROP COLUMN old_email;

### Example 3: Code Review Skill

**File:** `.opencode/skills/code-review/SKILL.md`

```markdown
---
name: code-review
description: "Comprehensive code review checklist"
---

# Code Review Skill

## Review Levels

### Level 1: Correctness
- Logic errors
- Edge case handling
- Error handling completeness
- Resource cleanup (files, connections, locks)

### Level 2: Quality
- Readability (naming, structure)
- Maintainability (complexity, coupling)
- Test coverage (critical paths 100%)
- Documentation (complex logic explained)

### Level 3: Architecture
- SOLID principles
- Design patterns (appropriate vs over-engineered)
- Separation of concerns
- Future extensibility

### Level 4: Context
- Follows project conventions
- Consistent with existing patterns
- Appropriate abstractions
- Security implications

## Output Format

### ✅ Approved
**Strengths:**
- List good patterns
- Praise excellent decisions

**Minor Suggestions:**
- Optional improvements
- Future considerations

### ⚠️ Needs Changes
**Required Changes:**
1. Issue with severity and location
2. Issue with proposed fix

**Explanation:**
- Why each change matters
- What could go wrong

### ❌ Requires Major Revision
**Critical Issues:**
- Architectural problems
- Security vulnerabilities
- Fundamental approach flaws

**Recommendation:**
- How to restructure
- Alternative approaches