What problem does it solve? Autonomous agent loops running 24/7 tend to fake PASS results, silently burn budget, drift from committed configuration, and contradict their own reports. This Skill enforces thirteen operational laws so continuously scheduled multi-workflow systems produce verifiable receipts, deterministic state transitions, and auditable cost attribution instead of unverifiable prose. ## Core Features & Use Cases - Thirteen governing laws (L1–L13): single reconciler authority, external-only verification, verifier freeze, fail-closed sinks, ROI throttling, drift detection, and deterministic replayable loops. - Receipt-driven auditing: adversarial checklist for reviewing agent reports — timestamp math, self-verification detection, count/ID validation, and cost sanity checks against defined JSON receipt schemas. - Kaizen maintenance pipeline: converts failures, drift receipts, and throttle events into an ROI-ranked improvement queue with machine-safe versus founder-gated classification and auto-rollback. - Use Case: A founder pastes an agent report claiming a deploy passed. The Skill flags that verification ran 30 seconds after publish (under the 60-second minimum), that the agent verified its own work, and that the cycle receipt lacks metered cost — rejecting the PASS and demanding external-runner evidence. ## Quick Start Ask the assistant to audit this agent worker report against the governed-autorun laws and identify any verification or cost violations.
