What problem does it solve? Product definitions and codebases drift apart over time: features get built without user stories, stories describe behavior the code no longer implements, and acceptance criteria silently diverge from reality. This Skill systematically compares a DefProd product area's user stories and acceptance criteria against the actual code, surfacing drift in both directions. ## Core Features & Use Cases - Bidirectional drift detection: Finds stories with no implementation, partially implemented acceptance criteria, implemented features with no story, and stories that diverge from the code. - Structured four-table report: Saves a dated discrepancy report into your e2e or docs directory, with per-row recommendations (update definition, fix code, remove, or create story). - Mandatory walkthrough workflow: Walks every report row with the user via confirmation prompts, updating the saved file row-by-row and appending a walkthrough log. - Use Case: After a sprint of rapid feature work, run the analysis on your BILLING area to discover that two new endpoints lack user stories and one acceptance criterion's validation rule no longer matches the code, then confirm each recommended fix before handing off to the fix workflow. ## Quick Start Ask the agent to run /defprod-analyze-discrepancies for your product name and area key, for example "analyze discrepancies for the AUTH area of my product".
