多Agent编码团队任务委派与冲突解决规范

You are a senior engineering manager designing collaboration protocols for a multi-agent coding team. I have multiple AI coding agents working on the same codebase simultaneously.

My Team Setup

• Number of agents: [e.g., 3-5]
• Platform: [e.g., Multica / Claude Code instances / Codex workers]
• Repository: [e.g., monorepo / microservices / single app]
• Languages: [e.g., TypeScript, Python]
• CI/CD: [e.g., GitHub Actions / GitLab CI]

Generate For Me

1. Task Decomposition Rules

• How to break a feature into agent-assignable units
• Dependency graph template
• Maximum task granularity guidelines

2. File Ownership Protocol

• File locking / reservation system design
• Directory-level ownership mapping
• Shared file editing rules (e.g., package.json, schema files)

3. Conflict Prevention

• Branch naming convention per agent
• Pre-merge checks each agent must run
• Automated conflict detection before PR

4. Communication Protocol

• Inter-agent status update format
• Blocking dependency notification template
• Completion/handoff message format

5. Quality Gates

• Per-agent PR checklist
• Integration test requirements before merge
• Code review assignment rules

Output as a TEAM_PROTOCOL.md file ready to drop into the repo root.