An agentic software development methodology for AI coding agents, ported to OpenClaw's native skill format.
Based on obra/superpowers.
| Skill | Description |
|---|---|
| brainstorming | Collaborative design before implementation |
| writing-plans | Create implementation plans from designs |
| executing-plans | Execute plans with review checkpoints |
| subagent-driven-development | Execute plans using parallel subagents |
| dispatching-parallel-agents | Launch independent tasks in parallel |
| test-driven-development | RED-GREEN-REFACTOR cycle enforcement |
| systematic-debugging | 4-phase root cause analysis |
| verification-before-completion | Evidence before assertions |
| code-review | Requesting and receiving code reviews |
| using-git-worktrees | Isolated workspace management |
| finishing-a-development-branch | Branch completion and integration |
| writing-skills | Creating new skills |
| using-superpowers | Skill discovery and usage guide |
Copy skills and hook to your OpenClaw workspace:
cp -r skills/* ~/.openclaw/skills/
cp -r hooks/openclaw ~/.openclaw/hooks/superpowers
openclaw hooks enable superpowersopenclaw status # Should show skills loaded
openclaw hooks list # Should show superpowers hookThe 7-step superpowers workflow:
- Brainstorm — Design before implementation
- Plan — Write detailed implementation plan
- Execute — Implement plan with subagents or batches
- Test — TDD throughout (RED-GREEN-REFACTOR)
- Debug — Systematic root cause analysis when needed
- Review — Code review after each task
- Finish — Branch completion and integration
MIT