Cursor rule

1---
2description: "Agent orchestration: available agents, parallel execution, multi-perspective analysis"
3alwaysApply: true
4---
5# Agent Orchestration
6 
7## Available Agents
8 
9Located in `~/.claude/agents/`:
10 
11| Agent | Purpose | When to Use |
12|-------|---------|-------------|
13| planner | Implementation planning | Complex features, refactoring |
14| architect | System design | Architectural decisions |
15| tdd-guide | Test-driven development | New features, bug fixes |
16| code-reviewer | Code review | After writing code |
17| security-reviewer | Security analysis | Before commits |
18| build-error-resolver | Fix build errors | When build fails |
19| e2e-runner | E2E testing | Critical user flows |
20| refactor-cleaner | Dead code cleanup | Code maintenance |
21| doc-updater | Documentation | Updating docs |
22 
23## Immediate Agent Usage
24 
25No user prompt needed:
261. Complex feature requests - Use **planner** agent
272. Code just written/modified - Use **code-reviewer** agent
283. Bug fix or new feature - Use **tdd-guide** agent
294. Architectural decision - Use **architect** agent
30 
31## Parallel Task Execution
32 
33ALWAYS use parallel Task execution for independent operations:
34 
35```markdown
36# GOOD: Parallel execution
37Launch 3 agents in parallel:
381. Agent 1: Security analysis of auth module
392. Agent 2: Performance review of cache system
403. Agent 3: Type checking of utilities
41 
42# BAD: Sequential when unnecessary
43First agent 1, then agent 2, then agent 3
44```
45 
46## Multi-Perspective Analysis
47 
48For complex problems, use split role sub-agents:
49- Factual reviewer
50- Senior engineer
51- Security expert
52- Consistency reviewer
53- Redundancy checker
54