showdar-understandlisted

# Showdar Understand ## Purpose - Build a trustworthy mental model of a repository before changing it. - Identify runtime entrypoints, ownership boundaries, state/data flow, dependencies, build/test systems, and platform surfaces. - Separate facts observed in files from architectural inferences. - Reduce random file reading by following high-signal configuration and representative runtime paths. - Read `references/repository-analysis.md` for the default inspection sequence. - Use `references/architecture-signals.md` when directory names and actual dependency direction disagree. - Use `references/dependency-analysis.md` when the task crosses packages, native modules, persistence, queues, or external services. ## When to use - Entering a repository or subsystem you do not already understand. - Before planning a feature whose change surface is unclear. - Before a refactor or migration that may cross package/module boundaries. - When the user asks for an architecture overview, dependency map, ownership map, or “where should this change go?” - When a bug appears to cross layers and you need a minimal map before debugging. - When agent context was compacted and repository structure must be re-established. ## When not to use - The task is a tiny, already-localized edit and the exact owning file/contract is known. - The user only asks a conceptual question unrelated to the repository. - A production incident requires immediate evidence collection from runtime systems; use the r