spec-review

To check codebase coverage, the skill parses a PRD file or URL into atomic requirements, scans source files, and classifies each requirement as implemented, partial, missing, or extra. This generates a detailed coverage report highlighting gaps between the spec and the actual code.

Spec drift detection identifies added, removed, or altered requirements by comparing new specifications with previously stored versions. It continuously validates implementation against the spec to prevent unintended functionality and scope creep over time.

Validating implementation against a PRD involves decomposing the document into testable requirements, scanning the codebase using search hints, and logging decisions. It produces actionable reports with visual coverage percentages and gap lists for future reference.

Spec review requires read, grep, glob, and write tools to function. These tools are necessary to parse specifications, scan source files within the codebase, and generate the final coverage report with drift analysis.

Yes, you can provide a PRD via a file path or a URL. The skill reads the baseline specification from the provided source, decomposes it into atomic requirements, and proceeds with codebase scanning and coverage classification.

Preventing scope creep requires continuously validating implementation against the spec to detect extra or changed requirements. By comparing current code against a baseline PRD and generating a coverage report, teams can identify unintended functionality and maintain strict alignment.