coordinator-dx
Coordinator playbook for running multi‑repo, multi‑VM work in parallel without relying on humans copy/pasting long checklists.
📖 中文导读:Coordinator playbook for running multi‑repo, multi‑VM work in parallel without relying on humans copy/pasting long checklists. 以下正文为该技能的原始 SKILL.md 内容(英文),上方安装命令可直接使用。
name: coordinator-dx description: Coordinator playbook for running multi‑repo, multi‑VM work in parallel without relying on humans copy/pasting long checklists.
coordinator-dx
Coordinator playbook for running multi‑repo, multi‑VM work in parallel without relying on humans copy/pasting long checklists.
Key conventions:
- Each VM/agent sets
AGENT_NAME=<vm>-<tool>(e.g.macmini-codex,epyc6-claude-code,homedesktop-wsl-gemini) - For in-repo coordination,
thread_idshould equal the repo’s local Beads issue id (no cross-repo renames). - Always start a session with
bash scripts/cli/dx_doctor.shand follow the warnings (soft, advisory).
Recommended coordinator flow:
- Assign work by repo (prime-radiant-ai / affordabot / llm-common / agent-skills).
- Require agents to report:
git status --porcelaingit branch --show-current- output of
bash scripts/cli/dx_doctor.sh
- Prevent collisions:
- one repo per VM by default
- prefer small PRs
- Enforce “automation not checklists”:
- if a step is repeated, codify it in
dx_doctor.shor shared skills.
- if a step is repeated, codify it in