vidhi-onboardlisted

# Sutra Onboarding Generate a `.sutra/rules.toml` for a codebase by discovering its architecture, analyzing dependency patterns, and deriving constraints from what's actually there. The output is a reviewed, committed rules file — not a template. This is an interactive skill. You explore, analyze, present findings, and let the user shape the rules before writing. ## Preconditions - Sutra MCP server must be available (check for `sutra_status` tool). If not, stop and tell the user. - The codebase must be in a language sutra supports (Rust, Dart). If unsure, check with `sutra_help`. ## Entry The user provides a path to the workspace root, or you infer it from the current working directory. If the workspace has a `Cargo.toml` or `pubspec.yaml`, auto-detect the language. Otherwise ask. ## Process ### 1. Register and parse Check if the workspace is already registered: ``` sutra_status(path="/absolute/path/to/workspace") ``` If `status` is not `ready` or `is_stale` is true, trigger a parse: ``` sutra_parse(workspace="<name>") ``` Wait for parse to complete (re-check status). Report the file count, symbol count, and any parse errors. ### 2. Map the architecture Run these in parallel to build a picture of the codebase: ``` sutra_map(workspace="<name>", limit=40) sutra_components(workspace="<name>") sutra_deps(workspace="<name>") sutra_conventions(workspace="<name>", action="list") ``` Also explore the repo for architecture docs: - Look for `docs/adr/` or similar ADR