superpowers-verification

What This Skill Does

The superpowers-verification skill is a mandatory quality assurance protocol for OpenClaw agents, designed to eliminate the 'fake success' syndrome. It codifies the principle of 'evidence before assertions.' When an agent performs a task, this skill ensures that no claim of completion—whether it be a bug fix, a successful build, or a passed test—is made without executing a fresh, ground-truth verification command. By forcing the agent to identify the correct verification command, run it, and inspect the exit code or output logs before reporting back to the user, this skill effectively bridges the gap between 'code that looks correct' and 'code that actually works.' It mandates a 'Red-Green-Refactor' mindset, requiring proof of failure before success to ensure regression stability and total system integrity.

Installation

You can integrate this skill into your environment by running the following command in your terminal: clawhub install openclaw/skills/skills/axelhu/superpowers-verification

Use Cases

This skill is essential in any development lifecycle involving:

• Software Debugging: Ensuring a bug is truly gone by running the specific failing test case.
• Test-Driven Development (TDD): Validating that features meet all requirements through passing test suites.
• CI/CD Pipelines: Confirming builds or linting processes exit with code 0 before initiating a commit or deployment.
• Regression Testing: Protecting against code rot by enforcing that previous features still function after new modifications.
• Agent Orchestration: Ensuring delegated tasks from other agents were successfully completed before moving to the next sequence.

Example Prompts

1. "I've finished the authentication module integration, please run the superpowers-verification protocol to confirm all 50 tests pass before I push to production."
2. "Before you report the bug as fixed, perform the required verification check with the original test case and confirm the failure count is zero."
3. "Is the linter clean? Run the verification skill and show me the output snippet confirming no errors remain."

Tips & Limitations

To maximize the effectiveness of this skill, strictly avoid using 'probabilistic' language such as 'should,' 'looks good,' or 'probably.' The verification process requires an objective, binary outcome. Limitations include a slight increase in latency for individual tasks, as execution time is required for the verification phase. However, this cost is vastly outweighed by the savings in avoiding debugging cycles, broken builds, and the collapse of trust between the agent and the human operator. Always treat the verification checklist as non-negotiable.