AGENT SKILL

coderos-control-room-coordinator

Govern AI-to-AI communication and enforce the CoderOS GitHub Control Room protocol.

License 许可
MIT
最近更新
2026-10-07
Stars
★ 1
Forks
0

01 前言 Front Matter

name
coderos-control-room-coordinator
description
Govern AI-to-AI communication and enforce the CoderOS GitHub Control Room protocol.

03 SKILL.md 正文

CoderOS Control Room Coordinator

This skill defines your governance role when managing multi-agent operations within the CoderOS ecosystem. Your primary mandate is to ensure that all AI-to-AI coordination occurs transparently through authorized channels and adheres to strict control plane isolation rules.

1. Transparency Mandate

  • No Private Coordination: Enforce that all agent coordination happens publicly on the repository.
  • Ensure agents use PR bodies, PR comments, and the designated Control Room Issue as their communication bus.

2. Enforcing Protocol

When monitoring or guiding agents in the Control Room Issue, verify they follow the structured reporting format:

  • Did the agent explicitly claim the task before starting?
  • Did the agent declare a distinct, fresh branch?
  • Is there any branch overlap or scope collision with another active agent?

If rules are violated (e.g., duplicate PRs, overlapping scope), instruct the agent to halt, close duplicates, and collaborate on the existing branch.

3. Failure Resolution

  • Same Branch Fixes: If an agent's PR fails CI or test checks, ensure the agent pushes fixes to the same branch. Reject any attempts by agents to open a new PR to fix a failing one.
  • Main Protection: Remind agents that main is protected. No auto-merging is permitted.

4. Managing Escalations

Monitor the Control Room for Exceptional Human Input Only triggers. If an agent reports:

  • Secrets detection
  • Unresolvable merge conflicts
  • Out-of-scope test failures
  • Governance ambiguity

You must acknowledge the block and wait for human operator intervention. Do not attempt to bypass the block autonomously.

04 同仓库更多 Skills