A collection of 14 OpenClaw skills that enforce disciplined, verifiable agent workflows.
| Skill | When to Use |
|---|---|
| ------- | ------------ |
| brainstorming | Before creative or feature-design work |
| dispatching-parallel-agents | 2+ independent tasks that can run in parallel |
| executing-plans | Execute a written plan with checkpointed reviews |
| finishing-a-development-branch | After implementation, choose merge/PR/cleanup path |
| receiving-code-review | Processing review feedback, especially ambiguous comments |
| requesting-code-review | Before merge to verify quality and requirements |
| subagent-driven-development | Execute plans via independent subagent tasks |
| systematic-debugging | Bugs and failures — root cause before fixes |
| test-driven-development | Before writing implementation code |
| using-git-worktrees | Feature work that needs workspace isolation |
| using-superpowers | Session start — identify the right skill first |
| verification-before-completion | Before any "done" claim — prove it works |
| writing-plans | Multi-step tasks that need concrete execution plans |
| writing-skills | Creating, editing, or validating skills |
npx clawhub@latest install superclaw
Stop guessing. Stop claiming "done" without evidence. Stop debugging by random changes. These skills enforce the patterns that separate reliable agent work from chaos.
共 1 个版本
暂无安全检测报告