What problem does it solve? Completed spec directories in a spec-kit project accumulate valuable architectural decisions and implementation patterns that are lost unless manually distilled into durable documentation. This Skill automates that distillation, moving decisions into ADRs, system knowledge, and coding guidelines. ## Core Features & Use Cases - Multi-spec batch processing: Resolve and process several spec directories in one run, with per-spec validation and extraction-status checks. - Classified extraction: Parse research.md, spec.md, data-model.md, and contracts/ to classify content into ADRs, knowledge updates, and prescriptive guidelines. - Interactive review gate: Present a numbered extraction plan and write nothing until the user approves, skips, edits, or groups items. - Use Case: After finishing a feature spec like specs/042-graphql-name-lookup, run the extraction to generate numbered MADR files in dev/adr/, update dev/knowledge/ and dev/guidelines/, then archive the spec under specs/archive/ with an EXTRACTED.md record. ## Quick Start Extract all durable knowledge, guidelines, and ADRs from the completed spec directory specs/042-graphql-name-lookup into the dev documentation folders.
