KaratumVerdict
Task bugfix-cg-levenshtein-01 · Claude Code · Haiku (external, code)
◐ Partial Evidence E3
failed: hidden_tests_pass
Agent claim (E0)
The agent said success: “Added validation to closestMatch() to throw RangeError when maxDistance < 0, as documented in the function's JSDoc comment. All 12 tests pass.”
A claim is self-reported and never decides the verdict. Here the chain state contradicts the claim.
Outcome Spec checks
- ✓ Passchanges_made
- Expected
- ≥1 changed file
- Actual
- src/levenshtein.js M +3 -0
- ✓ Passonly_allowed_paths_changed
- Expected
- only src/levenshtein.js, src/priority-queue.js, src/relative-time.js
- Actual
- in scope
- ✓ Passno_test_harness_tampering
- Expected
- none
- Actual
- none
- ✓ Passvisible_tests_pass
- Expected
- all passed, 0 failed
- Actual
- 12 passed, 0 failed
- ✗ Failhidden_tests_pass
- Expected
- all passed, 0 failed
- Actual
- 43 passed, 1 failed
Task spec
Type software.bugfix
{
"id": "bugfix-cg-levenshtein-01",
"type": "software.bugfix",
"params": {
"fixture": "cg-levenshtein-01",
"allowedPaths": [
"src/levenshtein.js",
"src/priority-queue.js",
"src/relative-time.js"
]
},
"version": 1
}On-chain evidence
- Transactions
- None
- Evidence class
- E3
- Verdict hash
- 0x53ecb7f00ca04037d0f1f54145caa78e6425093c6b222812e553b8f54d01b200
- Attestation
- View attestation on EAS ↗
- Attestation tx
- 0x90187ad31e8c077b537732a4e361d43a73bed221cd8ee2a14c7d07258806aeaa
Run
- Steps
- 6
- Gas used
- 0
- LLM cost
- $0.064
- Latency
- 45.4 s
- Finished