Import a codebase into CodeSpring

Turn a real repo into an accurate CodeSpring model: core features → sub-features → notes → PRDs → tasks. Reads code and git; writes only to CodeSpring — never modifies the codebase.

Shared knowledge, read it: codespring skill references/project-state.md (enter-from-anywhere, the five questions), references/auditing-and-fixing.md (audit → verdict → map → tasks, and why the map shows the target), references/analyze-codebase.md, references/mindmap-structure.md (the exact node/edge schema), references/prd-management.md, references/pitfalls.md. Then this skill's references/target-map.md (the map rules — the node model, the corrected target, and how to keep it small).

Language- and stack-agnostic. Any named specific here or in the references is a labelled worked example, never an assumption about the repo in front of you.

Script paths are relative to this skill's own folder; the skills install side by side, so ../codespring/scripts/… reaches the core skill's scripts. Resolve them to absolute paths once at the start — your working directory will be the user's repo.

0. Detect where the project already is

Run the state-detection scripts rather than counting by hand (the codespring skill's scripts/, installed alongside this one):

bash ../codespring/scripts/fetch-project.sh --out /tmp/cs-state   # read-only snapshot; prints the projectId
node ../codespring/scripts/state.mjs       /tmp/cs-state          # summary line + booleans and counts
node ../codespring/scripts/check-map.mjs   /tmp/cs-state          # machine-checkable map faults, before you add to them