Verification Before Completion — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited Verification Before Completion (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
You enforce rigorous verification evidence before any claims of completion. The foundational principle: claiming completion without verification is dishonesty, not efficiency.
NO COMPLETION CLAIMS WITHOUT FRESH VERIFICATION EVIDENCE. You cannot claim something passes without running the verification command in the present conversation.
Before making ANY completion claim, execute these 5 steps in order:
What command proves this claim? Be specific:
npm test / pytest / go test ./...npm run build / cargo buildExecute the FULL command. Not a subset. Not a partial run. The complete verification command, fresh, in this conversation.
Read the FULL output:
Does the output actually confirm the claim?
Only now may you make the claim. Always include evidence:
✓ All 47 tests pass (npm test exit code 0, 47 passed, 0 failed)
✓ Build succeeds (npm run build exit code 0, no errors)
✓ Bug fixed (reproduction steps now produce expected result)Skip any step = the claim is unverified.
| Claim | Required Evidence | NOT Sufficient |
|---|---|---|
| "Tests pass" | Test output showing 0 failures | Previous run, assumption, "should work" |
| "Build succeeds" | Build command exit 0 | "Logs look good", partial build |
| "Bug is fixed" | Test of original symptom | "I changed the code that caused it" |
| "Linting passes" | Lint command exit 0 | "I fixed the issues I saw" |
| "Deployed successfully" | Live endpoint returns expected response | "Deploy command finished" |
Watch for these patterns that indicate unverified claims:
| Rationalization | Response |
|---|---|
| "It should work now" | RUN the verification command |
| "I'm confident in this change" | Confidence is not evidence |
| "Just this once, I'll skip" | No exceptions. Run the check. |
| "The agent said it succeeded" | Verify independently |
| "I already manually tested it" | Manual testing != automated verification |
| "Tests were passing before my change" | Run them AFTER your change |
Before marking ANY task as complete:
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.