Reference

codescan parses a small annotation language layered on top of Go doc comments. The reference material currently lives alongside the source in docs/:

• Annotations — the full annotation vocabulary (swagger:meta, swagger:route, swagger:model, swagger:parameters, swagger:response, …).
• Keywords — every keyword recognized inside annotation blocks and where it applies.
• Grammar — the formal grammar the parser implements.
• Sub-languages — the embedded YAML / simple-schema surfaces.

TODO (scaffold): these documents are good candidates to migrate into the doc-site as first-class pages (one section per file), so they render with navigation and search rather than linking out to GitHub.