What problem does it solve? Keeping a deterministic simulation engine's calibration corpus, grade anchors, and parity documentation in sync is a repetitive, error-prone chore. This Skill automates the repeatable SimQ audit cycle so drift is detected, classified, and either resolved as a chore commit or escalated into a ticket without manual orchestration. ## Core Features & Use Cases - Calibration Recalibration: Runs make simq-full-audit in fast, full, or slow modes to refresh the calibration corpus and capture REGRESS rows, pytest counts, uncovered anchor keys, and parity ledger candidates. - Drift Classification & Anchor Updates: Classifies each flagged item as EXPECTED_DRIFT, REGRESSION, DA_NEEDED, or NO_ACTION, then updates grade_anchors.json and anchor key lists only for expected drift, re-running targeted pytest subsets as a gate. - Docs & Parity Sync with Governance Branching: Updates evaluation matrix, integration, and parity ledger docs, then either proposes a chore commit for clean passes or spawns a ticket via the ticket-scoper role for regressions. - Use Case: After merging simulation changes, invoke the audit in full mode to verify no regressions crept in; if anchors drifted intentionally, they are refreshed and a chore commit is suggested, while genuine regressions produce a ticket handed off to /implement-ticket. ## Quick Start Ask the AI to run the SimQ audit in full mode for the frontier_extended and wilderness_survival worlds and report whether any regressions were found.
