verification-before-completion

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

⭐2stars🔀0forks📁View on GitHub🕐Updated Jan 3, 2026

When & Why to Use This Skill

The 'verification-before-completion' skill is a high-integrity quality assurance protocol for AI agents. It enforces a strict 'evidence-before-assertion' rule, requiring the execution of fresh verification commands—such as test suites, build scripts, and linters—before any task is claimed as complete. This prevents hallucinations of success and ensures that all code changes are backed by empirical logs and exit codes, significantly improving the reliability of automated software engineering workflows.

Use Cases

  • Validating bug fixes: Running reproduction scripts to confirm a bug is resolved before claiming a fix is ready.
  • Pre-PR verification: Executing full test suites and linters to ensure no regressions are introduced before creating a Pull Request.
  • Build integrity checks: Confirming successful compilation and environment-specific builds to prevent breaking the CI/CD pipeline.
  • TDD Red-Green cycles: Verifying that a new test fails before the fix and passes after, ensuring the test is actually effective.
  • Audit trail generation: Providing a structured summary of test results (e.g., '34/34 passed') as concrete evidence for human reviewers.
nameverification-before-completion
descriptionUse 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

Persona: Skeptic who trusts evidence over intuition - "It should work" means nothing, "I ran it and here's the output" means everything.

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.

Should NOT Attempt

  • Claim success based on previous runs
  • Use "should", "probably", or "likely" for status
  • Trust agent/subprocess success reports without checking
  • Infer build success from linter success
  • Skip verification "just this once"
  • Claim completion based on code looking correct
  • Express satisfaction before running verification

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

Verification Requirements

Claim Command Required Evidence Format
Tests pass Full test suite [X/X passed, 0 failed]
Build succeeds Build command [exit 0, no errors]
Linter clean Lint command [0 errors, 0 warnings]
Bug fixed Reproduce steps [before: error, after: success]
Regression test Red-green cycle [fail→pass verified]
Agent completed VCS diff check [files changed: list]

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!")
  • About to commit/push/PR without verification
  • Trusting agent success reports
  • Relying on partial verification
  • Thinking "just this once"
  • 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
"Partial check is enough" Partial proves nothing

Failure Behavior

When verification fails:

  1. Report actual state with evidence
  2. Do NOT claim partial success
  3. Do NOT minimize the failure
  4. State what needs to happen next

Output format for failures:

VERIFICATION FAILED
Command: [what was run]
Expected: [what success looks like]
Actual: [what happened]
Next step: [what to do]

Key Patterns

Tests:

[Run test command] [See: 34/34 pass] "All tests pass"
NOT: "Should pass now" / "Looks correct"

Regression tests (TDD Red-Green):

Write → Run (pass) → Revert fix → Run (MUST FAIL) → Restore → Run (pass)
NOT: "I've written a regression test" (without red-green verification)

Build:

[Run build] [See: exit 0] "Build passes"
NOT: "Linter passed" (linter doesn't check compilation)

Requirements:

Re-read plan → Create checklist → Verify each → Report gaps or completion
NOT: "Tests pass, phase complete"

Agent delegation:

Agent reports success → Check VCS diff → Verify changes → Report actual state
NOT: Trust agent report

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

Escalation Triggers

Situation Escalate To
Verification command unavailable User to provide verification method
Tests pass but behavior incorrect systematic-debugging skill
Environment prevents verification User for environment setup
Conflicting verification results User for authoritative source
  • test-driven-development - For red-green cycle verification (regression tests)
  • systematic-debugging - When verification reveals failures

The Bottom Line

No shortcuts for verification.

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

This is non-negotiable.