What problem does it solve?
It prevents “looks done” outcomes by turning the final verification of an AI change into a structured, spec-driven check that confirms scope boundaries, acceptance criteria, and test evidence.
Core Features & Use Cases
- Completeness check vs spec: Ensures all P0/mandatory acceptance criteria and required tasks are actually satisfied or explicitly marked as not done.
- Correctness and drift detection: Confirms the implementation matches spec/plan and flags stale docs or unplanned scope creep instead of silently rewriting requirements.
- Scope boundary enforcement: Verifies changed files comply with
scope.md (freely editable vs requires confirmation vs forbidden) and blocks completion on violations.
- Evidence-driven verification: Runs the checks prioritized by
verification.md and records automated/manual results, gaps, risks, and the final decision.
Quick Start
Ask an AI agent to run workflow-kit-verify and update workspace verification.md for the change.