deep-analyze

You can extract architectural decisions from documentation by running a deep codebase analysis that pattern matches rules, decisions, tech stack, workflows, and conventions within your `docs/` directory, turning unstructured text into structured context entries.

Analyzing a docs directory for onboarding is best handled by deep documentation analysis, which categorizes domain entities, workflows, and established rules to quickly provide new developers with a clear understanding of project architecture and technology choices.

Pattern matching for documentation extraction works by scanning text files in the `docs/` directory to identify and categorize key information like rules, decisions, tech stack, workflows, domain entities, and conventions into structured context entries.

Yes, automating context entry creation from documentation requires executing the `kodo analyze --deep --auto` command, which triggers the deep codebase analysis to process your `docs/` directory and extract structured learnings.

Deep analysis is specifically designed to process documentation files located within the `docs/` directory, meaning unstructured text stored in other locations or file formats outside this structure is not supported by the current pattern matching workflow.

The specific categories of information extracted from project documentation include rules, decisions, tech stack specifications, workflows, domain entities, and conventions, all transformed into structured context entries for actionable knowledge.