| name | verification |
| description | Enforce evidence-based verification before claiming any task is complete |
Methodology from obra/superpowers (MIT)
Verification
Iron law: no claims without fresh evidence.
The Verification Gate
Before you say "done", "works", "fixed", or "verified", you MUST:
- Run the relevant command (test, build, lint, curl, etc.).
- Read the full output -- not just the exit code.
- Confirm the output matches the expected result.
- Only then claim completion.
If you cannot run a verification command, say so explicitly. Never assume.
What Counts as Verification
| Claim | Minimum Evidence |
|---|
| "Tests pass" | Paste or reference the test runner output showing green. |
| "Build succeeds" | Show the build command output with zero errors. |
| "Bug is fixed" | Show the reproduction case now producing correct output. |
| "File updated" | Read the file back and confirm the expected content. |
| "Service is running" | Hit the health endpoint and show the response. |
Workflow
- Finish your change.
- Decide which claims you are about to make.
- For each claim, run the matching verification step.
- If any step fails, fix and re-verify. Do NOT skip ahead.
- Report results with evidence (command + output).
Anti-patterns to Avoid
- Saying "should work" without running anything.
- Running a command but not reading its output.
- Verifying one thing and claiming another.
- Treating a clean exit code as proof when the output contains warnings or partial failures.
- Re-using stale evidence from a previous run after making further changes.