plan-polish

To convert a rough plan into a technical design document, you provide intake data and research dossiers to generate an implementation-grade design file with architecture outlines, data structures, and validation requirements tailored for autonomous agent execution.

Generating an implementation-grade technical design document requires an INTAKE.md file, a rough plan, and a RESEARCH_DOSSIER.md to apply a track-aware schema that drives the detailed architecture and component outlines.

A track-aware schema determines the depth of your technical design documentation by generating either a Lite schema for Track M or a Full schema for Track L, ensuring the output matches the project's specific implementation requirements.

Yes, you can use rough intake data to produce architecture and component outlines by processing it into a TECHNICAL_DESIGN.md file that includes thorough verification notes and edge-case handling specifically structured for autonomous coding agents.

The best way to structure technical design documents for autonomous agent execution is to use a track-aware schema that provides detailed data structures, architecture sections, and validation requirements, minimizing execution ambiguity.

You need a Full track schema for Track L projects and a Lite track schema for Track M projects, with the specific track determined directly from the INTAKE.md data to match the required implementation depth.