source-driven-development

Validate code against official documentation using a four-step process: DETECT project versions, FETCH current docs, IMPLEMENT patterns, and CITE sources. This grounds framework decisions in verifiable, up-to-date official documentation rather than stale training data.

AI training data becomes stale as APIs change, patterns are deprecated, and new best practices emerge. Grounding code in official documentation via a structured DETECT, FETCH, IMPLEMENT, and CITE workflow ensures framework decisions use current, trustworthy guidance.

Generate cited boilerplate by providing project dependency files so the skill can detect versions, fetch official documentation, implement patterns, and cite sources. This process ensures generated code matches current framework best practices with structured source verification.

Yes, documentation verification applies during architecture decisions, code reviews, and boilerplate generation across modern tech stacks. It uses an MCP-based lookup chain preferring Active MCPs and official docs before web fetches to validate framework decisions.

Verify multi-API changes during code reviews using a Documentation Verification Table with explicit UNVERIFIED flags. This governance mechanism enforces version-specific, current guidance by structuring citations and flagging unverified patterns.

Never rely on AI memory for framework patterns when APIs change, patterns are deprecated, or new best practices emerge. Enforce a detection, fetching, implementation, and citation workflow to ground every framework decision in current official documentation.