explain

Code explanations with cited local evidence are generated by reading repository context, .pipeline/config.yaml, and explain-spec references. This builds a concise evidence package that grounds explanations strictly in your project's available files and history.

Evidence-first code explanation generates clear, sourced descriptions for code, configs, and recent changes using only locally available files. This ensures explanations stay completely grounded in your project's actual repository context.

No, this read-only repository explainer does not modify your project files. All actions are strictly read-only by design, ensuring your code, configs, and repository history remain completely safe and unaltered during the evidence gathering process.

To get sourced explanations for recent git diffs, you ask for an explanation anchored in repository context. The tool reads local files, diffs, and docs to provide a clear, evidence-backed explanation of the recent changes.

Yes, you can use a subagent for reviewing code explanations. The tool supports optional subagent handoff for separation of concerns, allowing a secondary agent to review the evidence package and generated explanation.

The best way to explain project configs without external sources is to use an evidence-first tool that reads local files and config.yaml. It builds an explanation strictly from locally available evidence, ensuring no external dependencies are needed.