| name | prove |
| description | Verifies that a task is actually done by running its Verify step and showing the real output, no self-reporting. Use it after implementing a task, when you want to confirm it truly works and not just that the agent claims so. |
Prove: verify for real (no self-reporting)
Prove the task is done by running its Verify and showing the real evidence. This is the piece that keeps the agent from fooling itself. Ideally it runs in a different context/session than the one that implemented it.
Take the Verify step of the given task and run it:
- If it's a command: run it and paste the COMPLETE, real output (not a summary, not "all passed"). If it fails, show the error as-is.
- If it's a manual check: do the exact steps and report what you observed ("I did X, I saw Y"), including the failing path when it matters.
Hard rules:
- Do NOT declare success without pasted evidence.
- Do NOT invent command output or results. If you couldn't run it, say so and explain why.
- One task, one piece of evidence: if there's no evidence, the task is not done.
Final verdict: PASS / FAIL, with the evidence that backs it.