What problem does it solve? It provides a consistent, evidence-based knowledge base for judging quality gates QG-1 through QG-4 across the AgentDevFlow main workflow, so requirement definitions, Issue acceptance criteria, implementations, and final PR acceptance are evaluated against shared criteria instead of ad-hoc judgment. ## Core Features & Use Cases - QG-1 Definition Integrity: Validates requirement drafts and REQ/Decision files for classification correctness, measurable checkboxes, required sections, auto_gate completeness, and test strategy structure. - QG-2 Acceptance Criteria Coverage: Checks that Issue completion conditions traceably cover mandatory REQ requirements, numeric thresholds are reachable, and test strategies align with Epic scope exclusions. - QG-3 Implementation Deviation: Classifies implementation deviations as impl-bug, spec-bug, or scope-creep and recommends routing to case-run or case-revise. - QG-4 Final Acceptance: Verifies completion checkboxes, CI status, docs consistency, and bun test full-suite machine acceptance criteria before PR merge. - Use Case: When running /agentdev/case-run, the agent consults QG-3 criteria to detect deviations between the implementation and the Issue/REQ/Design before creating a PR, then reports findings for user decision. ## Quick Start Ask the agent to evaluate the current requirement draft against the QG-1 definition integrity gate and report pass, warn, or fail with evidence.