researching

Research codebase exploration by running git log, blame, and show commands to map structure, patterns, and dependencies. This produces structured artifacts—codebase.md and research.md—that document existing approaches and inform design decisions before you write new code.

Document patterns by mining git history for traceability and analyzing code structure. The research process captures architectural patterns, API usage, and dependency relationships in markdown artifacts that serve as reference for future work.

Yes. Research extracts external best practices and security or performance guidance alongside codebase analysis. Git history reveals how decisions were made and what trade-offs were accepted, informing safer design choices.

Research applies after requirements.md exists and planning requires understanding existing code patterns, APIs, or external documentation. This reduces rework risk by identifying constraints and opportunities before implementation begins.

Codebase research generates two markdown files: codebase.md documenting structure and patterns, and research.md containing findings and approach recommendations with trade-offs for multiple design options.

Yes. Research requires access to the repository, git log, blame, and show commands to trace code evolution and identify patterns. Task context helps focus the exploration on relevant areas of the codebase.