adlc-levelup-init

To discover patterns in a brownfield codebase for AI context directives, you can scan the repository to identify latent Context Directive Records (CDRs). This process analyzes the project structure and detects potential sub-systems to surface draft CDRs for team review.

Context Directive Records (CDRs) are patterns extracted from a codebase to feed team-ai-directives. You need to generate them when you want to establish structured AI context from an existing project, ensuring AI tools understand your established architectural conventions.

To scan your repository and extract draft CDRs, provide the brownfield repository path to initiate pattern discovery. The scan analyzes the project structure and outputs draft CDRs with the status "Discovered" to the designated drafts area for refinement.

Yes, you need a valid .specify project structure with a .specify/ directory to scan a codebase for CDRs. The tool outputs the discovered draft CDRs directly to {REPO_ROOT}/.specify/drafts/cdr.md for your team to review.

The best way to identify sub-systems in an existing codebase for AI directives is to run an automated pattern discovery scan. This detects potential sub-systems within the project structure and generates a concise summary with handoff options to proceed with your team.

Yes, you can use team-ai-directives with a codebase lacking CDRs by running a pattern discovery scan. It surfaces latent patterns from your existing repository and produces draft CDRs, providing a baseline for your team to review and refine into formal directives.