도구

Say build a harness for this project and it designs a domain-specific agent team, generating subagent definitions, companion skills and orchestration, validates them and keeps evolving them from feedback. Supports workflow orchestration, persistent agent teams and subagent delegation.

Tw93's lean pack of engineering-habit skills: /think to pressure-test plans, /hunt for root causes, /check to review diffs, /ui for frontend interfaces, /write to polish Chinese and English prose, and more (8 skills). Works with both Claude Code and Codex.

Shapes the agent's own replies rather than your documents: separates what changed from what was verified, cuts invented confidence, leads with the action and keeps load-bearing caveats. Manually invoked and stays on until you say stop; a script can enable it in CLAUDE.md or AGENTS.md.