explore-codebase

Explore codebases efficiently by spawning parallel subagents to map code areas concurrently. This approach divides exploration tasks across multiple workers, reducing total time and enabling structured discovery of patterns, dependencies, and module relationships across large repositories.

Yes. Parallel exploration maps critical code areas concurrently and returns YAML summaries of findings, gaps, and patterns. This surfaces module structure, interdependencies, and architectural context without sequential manual review.

Spawn multiple explorers to map distinct code areas in parallel, then consolidate YAML outputs to spot recurring patterns, shared dependencies, and architectural conventions. Parallel execution accelerates pattern discovery across large or unfamiliar codebases.

Parallel exploration deploys multiple subagents to investigate different code sections simultaneously, integrating Glob, Grep, Task, and Read tools. Sequential review examines files one at a time; parallel execution completes faster and produces structured YAML summaries of findings.

Structured YAML summaries documenting findings across explored areas, including code organization, identified patterns, module relationships, and exploration gaps. Output enables quick reference and downstream analysis without raw file review.

Parallel exploration is most effective for large or unfamiliar repositories where concurrent mapping of multiple code areas reduces total exploration time. Smaller codebases benefit less, but the approach scales efficiently with repository complexity.