
Verification Before Completion
Install this when you want your coding agent to prove tests, lint, and builds actually pass before it says a task is done or opens a PR.
Overview
Verification Before Completion is an agent skill most often used in Ship—review (also Ship—testing, Operate—iterate) that requires fresh command output before any claim that work is complete, fixed, or passing.
Install
npx skills add https://github.com/obra/superpowers --skill verification-before-completionWhat is this skill?
- Forces a fresh run of the exact command that proves each status claim (tests, lint, build) with full output read before
- Provides a repeatable gate: identify proof command → run → read exit code and failures → state claim only with cited evi
- Calls out common agent failures (stale runs, partial checks, log vibes) with a claim-vs-requirements table
Adoption & trust: 100k installs on skills.sh; 221k GitHub stars; 3/3 security scanners passed (skills.sh audits).
What problem does it solve?
You are ready to commit or open a PR but your agent may declare tests, lint, or builds passing without running verification in this session.
Who is it for?
Solo builders who ship through agent-assisted commits and PRs and want a hard stop against hallucinated green checks.
Skip if: Teams that rely entirely on hosted CI with merge gates and never trust local agent claims before merge anyway.
When should I use this skill?
About to claim work is complete, fixed, or passing, before committing or creating PRs—run verification commands and confirm output first; evidence before assertions always.
What do I get? / Deliverables
You get status statements backed by the actual command output, exit code, and failure counts from a full fresh run.
- Status updates that cite fresh command output, exit codes, and failure counts
- Blocked or revised completion claims when verification does not match the assertion
Recommended Skills
Journey fit
Spans multiple journey phases - primary shelf plus alternate fits below.
The skill is anchored at the ship boundary where solo builders merge work—commits and PRs—where unverified “all green” claims cause the most damage. Review is the canonical shelf because the documented trigger is explicitly before committing or creating PRs, after implementation but before integration is accepted.
Where it fits
Before marking a user story complete in your task list, the agent runs the project test suite and quotes zero failures from the live output.
Immediately before opening a PR, the agent runs lint and typecheck and states results with exit codes instead of assuming CI will catch issues.
When claiming regression coverage, the agent executes the full test command—not a subset—and reports failure counts explicitly.
After a production bug fix, the agent reproduces the failing test, reruns it, and shows passing output before declaring the issue resolved.
How it compares
Use as a procedural gate in the agent turn—not as a replacement for CI jobs or automated test runners in your pipeline.
Common Questions / FAQ
Who is verification-before-completion for?
It is for solo and indie builders using Claude Code, Cursor, Codex, or similar agents who need honest completion signals before commits and pull requests.
When should I use verification-before-completion?
Use it in Ship—review before commits and PRs, in Ship—testing when claiming tests or lint are clean, in Build—pm when marking a feature done, and in Operate—iterate when saying a bug is fixed—always with a fresh verification command in the same message.
Is verification-before-completion safe to install?
It is procedural guidance that instructs running local verification commands; review the Security Audits panel on this Prism page and treat shell execution like any other agent skill before enabling in sensitive repos.
SKILL.md
READMESKILL.md - Verification Before Completion
# Verification Before Completion ## Overview Claiming work is complete without verification is dishonesty, not efficiency. **Core principle:** Evidence before claims, always. **Violating the letter of this rule is violating the spirit of this rule.** ## The Iron Law ``` NO COMPLETION CLAIMS WITHOUT FRESH VERIFICATION EVIDENCE ``` If you haven't run the verification command in this message, you cannot claim it passes. ## The Gate Function ``` BEFORE claiming any status or expressing satisfaction: 1. IDENTIFY: What command proves this claim? 2. RUN: Execute the FULL command (fresh, complete) 3. READ: Full output, check exit code, count failures 4. VERIFY: Does output confirm the claim? - If NO: State actual status with evidence - If YES: State claim WITH evidence 5. ONLY THEN: Make the claim Skip any step = lying, not verifying ``` ## Common Failures | Claim | Requires | Not Sufficient | |-------|----------|----------------| | Tests pass | Test command output: 0 failures | Previous run, "should pass" | | Linter clean | Linter output: 0 errors | Partial check, extrapolation | | Build succeeds | Build command: exit 0 | Linter passing, logs look good | | Bug fixed | Test original symptom: passes | Code changed, assumed fixed | | Regression test works | Red-green cycle verified | Test passes once | | Agent completed | VCS diff shows changes | Agent reports "success" | | Requirements met | Line-by-line checklist | Tests passing | ## Red Flags - STOP - Using "should", "probably", "seems to" - Expressing satisfaction before verification ("Great!", "Perfect!", "Done!", etc.) - About to commit/push/PR without verification - Trusting agent success reports - Relying on partial verification - Thinking "just this once" - Tired and wanting work over - **ANY wording implying success without having run verification** ## Rationalization Prevention | Excuse | Reality | |--------|---------| | "Should work now" | RUN the verification | | "I'm confident" | Confidence ≠ evidence | | "Just this once" | No exceptions | | "Linter passed" | Linter ≠ compiler | | "Agent said success" | Verify independently | | "I'm tired" | Exhaustion ≠ excuse | | "Partial check is enough" | Partial proves nothing | | "Different words so rule doesn't apply" | Spirit over letter | ## Key Patterns **Tests:** ``` ✅ [Run test command] [See: 34/34 pass] "All tests pass" ❌ "Should pass now" / "Looks correct" ``` **Regression tests (TDD Red-Green):** ``` ✅ Write → Run (pass) → Revert fix → Run (MUST FAIL) → Restore → Run (pass) ❌ "I've written a regression test" (without red-green verification) ``` **Build:** ``` ✅ [Run build] [See: exit 0] "Build passes" ❌ "Linter passed" (linter doesn't check compilation) ``` **Requirements:** ``` ✅ Re-read plan → Create checklist → Verify each → Report gaps or completion ❌ "Tests pass, phase complete" ``` **Agent delegation:** ``` ✅ Agent reports success → Check VCS diff → Verify changes → Report actual state ❌ Trust agent report ``` ## Why This Matters From 24 failure memories: - your human partner said "I don't believe you" - trust broken - Undefined functions shipped - would crash - Missing requirements shipped - incomplete features - Time wasted on false completion → redirect → rework - Violates: "Honesty is a core value. If you lie, you'll be replaced." ## When To Apply **ALWAYS before:** - ANY variation of success/completion claims - ANY expression of satisfaction - ANY positive statement about work state - Committing, PR creation, task completion - Moving to next task - Delegating to agents **Rule applies to:** - Exact phrases - Paraphrases and synonyms - Implications of success - ANY communication suggesting completion/co