.cursorrules

# Cursor Rules for ContextGit

## This Project Uses contextgit for Requirements Traceability

This repository is managed with contextgit. The AI assistant MUST follow these workflows:

### When Making Changes to Requirements or Documentation

1. **BEFORE editing any file in `docs/`:**
   - Run `contextgit relevant-for-file <path>` to understand related requirements
   - Run `contextgit extract <ID>` to get precise context for specific requirements

2. **AFTER editing any file in `docs/`:**
   - Run `contextgit scan docs/ --recursive` to update the index
   - Run `contextgit status --stale` to check for broken links
   - If stale links found, update affected downstream items

### When Adding New Requirements

1. Generate a new ID: `contextgit next-id <type>`
   - Types: `business`, `system`, `architecture`, `code`, `test`, `decision`
2. Add YAML frontmatter to the Markdown file with the generated ID
3. Run `contextgit scan docs/ --recursive` to register it

### When a Requirement Changes

1. After editing, run `contextgit scan docs/ --recursive`
2. Check for stale downstream items: `contextgit status --stale --format json`
3. Update downstream items if needed
4. Mark as synchronized: `contextgit confirm <ID>`

### JSON Output for Parsing

Always use `--format json` when the output needs to be parsed:
```bash
contextgit show SR-010 --format json
contextgit status --stale --format json
contextgit extract SR-010 --format json
```

### Detection

This is a contextgit project if `.contextgit/config.yaml` exists.

## Project Info

- **Tool**: contextgit v1.0.1
- **Purpose**: Requirements traceability for LLM-assisted development
- **Index file**: `.contextgit/requirements_index.yaml`
- **Config file**: `.contextgit/config.yaml`