gsd-eval-review

AI evaluation coverage audit works by comparing executed phase evaluation data against the strategy specified in AI-SPEC.md. It identifies gaps where required evaluations were missed and generates a structured remediation plan to address them.

An AI evaluation remediation plan is a structured document detailing gaps found during an evaluation coverage audit. You need one after completing an AI development phase to verify all required evaluations were conducted and to outline corrective actions.

Generating an EVAL-REVIEW.md document requires accessing the AI phase's evaluation data and the AI-SPEC.md file. The audit process compares these inputs to produce a review containing a score, verdict, identified gaps, and a remediation plan.

Yes, AI evaluation audit requires an AI-SPEC.md file because it serves as the baseline strategy. The audit compares actual evaluation data against this specification to determine if the phase met the defined evaluation requirements.

The best way to verify AI phase evaluation effectiveness is to audit the evaluation coverage against the specified strategy in AI-SPEC.md. This generates a scored review document with a verdict and remediation plan for any missing evaluations.