What problem does it solve? Implemented changes often pass happy-path tests while hiding spec drift, missing edge cases, or data exposure risks. This Skill performs an independent adversarial review of an OpenSpec change after implementation and local verification, producing a severity-classified findings report and a clear archive verdict. ## Core Features & Use Cases - Spec and Task Alignment Audit: Compares implementation against proposal.md, design.md, and tasks.md to detect spec drift, incomplete tasks, and missing verification reports. - Ecommerce Domain Rule Checks: Verifies CustomerOrder vs SupplierOrder separation, ProductVariant as sellable unit, supplier data exposure, status separation, and purchase-time snapshot behavior. - Severity-Classified Verdict: Classifies findings as Blocker, Major, Minor, or Question and ends with a PASS, PASS WITH GAPS, or FAIL verdict plus an archiving recommendation. - Use Case: After finishing a checkout endpoint implementation, run this review from a fresh session to catch that supplier costs leak into a customer-facing API response before archiving the change. ## Quick Start Ask the AI to run an adversarial review on the OpenSpec change you just implemented and verified, before archiving it.
