What problem does it solve? Repeated verification runs often re-check the same healthy code while missing known blind spots, and failed checks get lost without a durable record. This Skill reads prior verification findings before testing and ensures failures are recorded in a structured, reproducible format. ## Core Features & Use Cases - Historical Bias: Reads .jfc/verification/index.md before substantial verification to bias tests toward previously fragile areas. - Mechanistic Reporting: Enforces a strict failure report format with exact command, observed output, expected output, and minimal reproduction. - Durable Persistence: Ending with VERDICT: FAIL or VERDICT: PARTIAL triggers the runtime to persist reports under .jfc/verification/reports/ and update the index. - Use Case: Before running a test suite on a large codebase, check the verification index, focus tests on modules that previously failed, and produce a reproducible failure report that future sessions can learn from. ## Quick Start Before running verification, read the verification index and bias your tests toward previously failing areas, then end any failure report with VERDICT: FAIL or VERDICT: PARTIAL.
