discover-acs

Discovering acceptance criteria involves a structured four-pass interview to clarify required behaviors in domain language before formalizing Given/When/Then specifications. This approach generates an acs.md file with numbered AC sections to guide subsequent spec creation.

Reverse-engineering acceptance criteria from existing code extracts candidate ACs from current specs, design documents, and implementations to identify gaps. It uses gap interviews to fill missing behaviors and produces a complete set of acceptance criteria for existing features.

You need to formalize acceptance criteria using domain language when a feature requires clear behavior definitions before locking in detailed Given/When/Then specifications. This prevents misalignment by capturing business rules in shared terminology before technical spec writing.

Yes, the acceptance criteria discovery process supports greenfield features by creating ACs from scratch through a structured interview. It also handles in-progress or completed features by reverse-engineering existing work to find missing criteria.

Acceptance criteria discovery requires checkpoint gating via dae_handoff.py and branch validation via dae_branch.py to enforce process consistency. These guardrails produce a deterministic handoff summary directing the next phase to formalize specs or reconcile consistency.

Extracting acceptance criteria before Given/When/Then specifications ensures teams clarify exact feature behaviors in domain language first. This prevents premature technical locking and produces verified exit criteria evidence for a smoother formal spec handoff.