gsd-eval-review

To generate a remediation plan for AI workflow evaluation gaps, run an audit of the executed AI phase against your AI-SPEC.md evaluation strategy. The process outputs an EVAL-REVIEW.md file containing a coverage score, verdict, identified gaps, and a detailed remediation plan.

AI phase evaluation coverage measures whether the evaluation strategy defined in AI-SPEC.md was fully implemented during an AI workflow. You need to audit this coverage immediately after an AI phase is completed to verify compliance and identify any evaluation gaps before moving forward.

You can check if your AI workflow evaluation strategy from AI-SPEC.md was followed by running an evaluation coverage audit on the completed phase. This audit cross-references your workflow execution data with AI-SPEC.md requirements to produce a compliance verdict and highlight missed evaluation criteria.

Yes, auditing AI workflow evaluation coverage requires access to evaluation reference materials and workflow execution data. The audit process uses YAML frontmatter and applies these reference inputs to ensure the executed phase complies with the evaluation strategy defined in AI-SPEC.md.

The best way to document AI phase evaluation gaps and remediation steps is to output a structured EVAL-REVIEW.md file. This file captures the evaluation coverage score, final verdict, specific gaps in the evaluation strategy, and a actionable remediation plan for necessary improvements.

Your AI phase evaluation coverage audit shows gaps when the executed workflow did not fully implement the evaluation strategy defined in AI-SPEC.md. The audit captures these deviations in an EVAL-REVIEW.md file, providing a score and a targeted remediation plan to resolve the missing evaluation criteria.