oleg494 Verification Before Completion
Overview
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
Evidence is keyed to the checked state: revision/snapshot, exact command, scope, environment, and run timestamp. Reuse existing evidence until an invalidation condition occurs (code change, failure, or unresolved concern); report that provenance instead of treating the chat turn as a clock. If no valid evidence exists for the current checked state, you cannot claim it passes.
The Gate Function
BEFORE claiming a verified result:
1. IDENTIFY: What command proves this claim?
2. RUN: Execute that command fresh and completely when no valid evidence
exists for the current checked state — "FULL" means no truncated or
partial run of the check that proves THIS claim, not every check the
repo happens to have
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. Reusing evidence that is still valid
for the checked state (same revision, no invalidating change/failure/concern)
is not skipping — but claiming without either is lying.
Scope is part of honesty, in both directions. Claiming a suite-wide green you did not run is a lie; re-running an unchanged check with no new changes, failures, or unresolved concerns is ceremony, not verification. Broaden when scope warrants: shared code touched, or a failure the targeted check exposed.
Common Failures
| Claim | Requires | Not Sufficient |
|---|---|---|
| Tests pass | Test command output: 0 failures on checked state | Unchecked assumption, "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 | Inspect changed artifacts and exercise the claimed behavior | Agent reports "success" |
| Requirements met | Line-by-line checklist | Tests passing |
Red Flags - Investigate Before Claiming Success
- Presenting "should", "probably" or "seems to" as proof, rather than explicitly labeled uncertainty
- 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):
PASS: Observe the bug fail in isolation before the fix; apply the fix; observe the same check pass. If the fix already exists, demonstrate the failure in a disposable pre-fix copy, not by reverting the user's working tree.
❌ "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:
PASS: Agent reports success -> inspect the changed artifacts -> verify the claimed behavior -> report actual state
❌ Trust agent report
When To Apply
Apply before a completion or correctness claim, commit, PR or integration. Planning, delegation and a task transition do not themselves require running unrelated checks. Reuse evidence only while it covers the checked state; an independent judge may rerun it to establish independent observation. Calibrated uncertainty is valid reporting, never a substitute for proof.
Source: obra/superpowers (MIT). Adapted for coding-kit: cross-references made local.