sartoris-digital/verification-before-completion

Verification Before Completion

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.

Violating the letter of this rule is violating the spirit of this rule.

Verification Tiers

Select the appropriate verification tier based on change scope:

Light Verification

When: <5 files changed, <100 lines, tests exist and cover the changes

Dispatch architect at fast tier:

subagent({
  agent: "architect",
  task: "MODE: verification\n\nClaimed: [what is being claimed]\n\nVerify:\n1. Run build/typecheck\n2. Confirm diagnostics clean\n3. Spot-check changed files match claimed scope",
  tier: "fast"
})

Evidence required: Build passes + diagnostics clean.

Standard Verification (Default)

When: Default for all tasks not matching light or thorough criteria

Dispatch architect at standard tier:

subagent({
  agent: "architect",
  task: "MODE: verification\n\nClaimed: [what is being claimed]\n\nVerify:\n1. Run full test suite\n2. Check all changed files match task scope\n3. Confirm no regressions\n4. Verify claimed features work",
  tier: "standard"
})

Evidence required: Build passes + full test suite passes + changes match scope.

Thorough Verification

When: >20 files changed, security-sensitive changes, architectural changes, or file patterns match **/auth/**, **/security/**, **/crypto/**

Dispatch architect at reasoning tier:

subagent({
  agent: "architect",
  task: "MODE: verification\n\nClaimed: [what is being claimed]\n\nVerify:\n1. Run full test suite\n2. Review all changed files in detail\n3. Check for security implications\n4. Verify architectural consistency\n5. Confirm no regressions across the full codebase",
  tier: "reasoning"
})

Evidence required: Full architect review + all tests pass + security check.

Automatic Tier Selection Hints

| File Pattern | Suggested Tier | |-------------|---------------| | **/auth/**, **/security/**, **/crypto/** | Thorough | | **/test/** only | Light | | *.md only | Light | | **/api/**, **/routes/** | Standard or Thorough | | >20 files | Thorough | | <5 files, <100 lines | Light | | Everything else | Standard |

The Gate Function

BEFORE claiming any status or expressing satisfaction:

1. IDENTIFY: What command proves this claim?
2. SELECT TIER: Light / Standard / Thorough based on change scope
3. DISPATCH: Architect verification at selected tier
4. READ: Full verification output
5. VERIFY: Does output confirm the claim?
   - If NO: State actual status with evidence
   - If YES: State claim WITH evidence
6. 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" | | Build succeeds | Build command: exit 0 | Linter passing, logs look good | | Bug fixed | Test original symptom: passes | Code changed, assumed fixed | | Requirements met | Line-by-line checklist | Tests passing | | Agent completed | VCS diff shows changes | Agent reports "success" |

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 without independent verification
• Relying on partial verification
• 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 | | "Agent said success" | Verify independently | | "Partial check is enough" | Partial proves nothing |

The Bottom Line

No shortcuts for verification.

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

This is non-negotiable.