What problem does it solve? Teams often accept "done" claims, PR descriptions, or completion assertions at face value, letting phantom artifacts and untested code slip through. This Skill provides a decorrelated, separate-harness verifier that re-runs the actual commands behind any claim and confirms cited artifacts and receipts truly exist before a verdict is trusted. ## Core Features & Use Cases - Falsifiable Claim Decomposition: Breaks a PR, commit, or "done" assertion into an explicit checklist of runnable commands and files that must exist. - Phantom Detection: Confirms every cited artifact and Runtime Truth Spine receipt exists on disk, flagging narrated-as-done work with missing evidence. - Structured Verdicts: Posts a PASS, FAIL, UNPROVEN, or BLOCKED verdict comment via gh pr comment and writes a JSON receipt under reports/loop_closure/oz_verify/. - Use Case: A teammate claims "Loop 5b closure shipped with receipt written." The verifier re-runs the closure command, finds the run green but the receipt file missing, and reports FAIL with the exact evidence row. ## Quick Start Ask the agent to independently verify a specific PR or done-claim by re-running its tests and confirming its cited artifacts and receipts exist.
