What problem does it solve? Before making a substantial code change or technical plan, teams need grounded evidence about how the existing repository actually works. This Skill traces local code, behavior, dependencies, seams, and tests, then produces a sanitized, reviewable Markdown research artifact instead of guesses. ## Core Features & Use Cases - Evidence Tracing: Follows symbols through callers, data flow, error paths, configuration, and tests, labeling each claim as fact, inference, assumption, or unknown with path and line references. - Approval-Gated Artifacts: Proposes the research document path and outline first, writes only after explicit approval, and never creates files implicitly. - Secret-Safe Output: Excludes credentials, tokens, .env contents, and sensitive logs, using typed placeholders like [REDACTED_TOKEN] only when necessary. - Use Case: Before refactoring a payment module, ask the assistant to trace its callers, dependencies, and test seams, then review the resulting research document before handing off to technical design. ## Quick Start Ask the assistant to use the repository-research skill to trace how a specific feature works and produce a reviewable research summary before any changes.
