levelup-init

To document implicit patterns in a brownfield codebase, you scan the repository to identify rules and generate Context Directive Records. This process surfaces implicit directives as structured draft entries for team collaboration and codebase analysis.

Context Directive Records are structured documents capturing implicit codebase patterns and organizing discoveries into contextual modules like rules, personas, and skills. They map multiple modules into sub-systems to guide codebase analysis and team AI directives.

Yes, a ready .specify directory in your project root is required. The scanning process uses this directory structure to output generated Context Directive Records directly into .specify/drafts/cdr.md for subsequent review and refinement.

You start automating context directive discovery by running a brownfield scan workflow against your repository. This automatically surfaces team AI directives and generates initial CDR drafts at .specify/drafts/cdr.md in your project root.

The best way to organize discovered rules and personas is mapping them into contextual modules during the CDR generation process. This structures discoveries into categories like rules, personas, examples, and skills for later subsystem mapping.