PluginBench
Skill
Pass
Audit score 90

verification-before-completion

obra/superpowers

Run verification commands and confirm output before claiming work is complete—evidence before assertions always.

What is verification-before-completion?

This skill enforces a discipline of running actual verification commands and reading their output before making any claims about work being complete, fixed, or passing. Use it whenever you're about to commit, create a PR, or declare a task done to prevent false completion claims and maintain honest communication.

  • Requires executing full verification commands fresh before any success claims
  • Enforces reading complete output and exit codes rather than relying on assumptions
  • Prevents common failures like trusting partial checks, agent reports, or "should work" reasoning
  • Provides a gate function: identify verification command → run it → read output → verify claim matches evidence
  • Catches red flags like expressing satisfaction before verification or committing without fresh evidence

How to install verification-before-completion

npx skills add https://github.com/obra/superpowers --skill verification-before-completion
Claude Code
Cursor
Windsurf
Cline

How to use verification-before-completion

  1. 1.Before claiming any status or expressing satisfaction, identify what command proves your claim
  2. 2.Execute the full verification command fresh (not from memory or previous runs)
  3. 3.Read the complete output, check the exit code, and count any failures or errors
  4. 4.Verify that the output actually confirms your claim—if not, state the actual status with evidence instead
  5. 5.Only after confirmation, make your claim and include the evidence in your communication

Use cases

Good for
  • Before committing code changes, run the full test suite and confirm zero failures in output
  • Before creating a PR, execute the build command and verify exit code 0 rather than assuming linter passing means build succeeds
  • Before claiming a bug is fixed, run the original failing test to confirm it now passes
  • Before delegating work to an agent, verify the agent's reported success by checking the actual VCS diff and running verification independently
  • Before moving to the next task, re-read requirements and verify each one against actual test output or checklist
Who it's for
  • Developers committing code or creating pull requests
  • Anyone delegating work to AI agents and needing to verify results
  • Teams that value honest communication and preventing false completion claims
  • Engineers working in safety-critical or high-stakes codebases where unverified claims cause real damage

verification-before-completion FAQ

What if I'm confident the code works without running verification?

Confidence is not evidence. You must run the verification command and read its output. No exceptions, even if you're tired or in a hurry.

Does a linter passing mean the build succeeds?

No. Linter passing and build success are different verification steps. You must run the actual build command and confirm exit code 0.

Can I trust an agent's report that it completed the task successfully?

No. Verify independently by checking the VCS diff and running verification commands yourself. Agent reports are not evidence.

What counts as a valid verification command?

A command that directly tests your claim: test suite for "tests pass", build command for "build succeeds", running the original failing test for "bug fixed". Partial checks do not count.

What if I've only partially verified the work?

Partial verification proves nothing. You must run the complete verification command and confirm the full output before making any claim.

Full instructions (SKILL.md)

Source of truth, from obra/superpowers.


name: verification-before-completion description: Use when about to claim work is complete, fixed, or passing, before committing or creating PRs - requires running verification commands and confirming output before making any success claims; evidence before assertions always

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

ClaimRequiresNot Sufficient
Tests passTest command output: 0 failuresPrevious run, "should pass"
Linter cleanLinter output: 0 errorsPartial check, extrapolation
Build succeedsBuild command: exit 0Linter passing, logs look good
Bug fixedTest original symptom: passesCode changed, assumed fixed
Regression test worksRed-green cycle verifiedTest passes once
Agent completedVCS diff shows changesAgent reports "success"
Requirements metLine-by-line checklistTests 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

ExcuseReality
"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/correctness

The Bottom Line

No shortcuts for verification.

Run the command. Read the output. THEN claim the result.

This is non-negotiable.