| name | ravenclaw-context |
| description | Load work context from Ravenclaw and prepare for task execution. Use at the start of a new session or when switching between tasks. |
This skill helps you quickly load and apply work context from the Ravenclaw system.
Steps
1. Load Work Context
First, get the full work context. If the Ravenclaw MCP server is available, use the get_work_context tool. Otherwise, run the CLI:
rc context
2. Review Active Work
Check what's currently in progress:
rc issue list --status=in_progress --format=table
Check what's blocked or ready to start:
rc issue list --status=todo --format=table
3. Check User Comments
If working on a specific issue, check for user comments/feedback:
rc comment list issue <issue-id>
4. Start Working
Pick an issue to work on and mark it:
rc issue start <key>
5. When Done
Update the issue and add any relevant wiki documentation:
rc issue done <key>
rc wiki write <topic-slug>
6. Epic Locks (Multi-Agent Safety)
Before starting work on an epic's issues, acquire a lock:
rc lock acquire <epic-id> --session $SESSION_ID --agent claude-code
Keep the lock alive during long work:
rc lock heartbeat <epic-id> --session $SESSION_ID
Release when done:
rc lock release <epic-id> --session $SESSION_ID
Or simply use rc issue start <key> — it auto-acquires the epic lock.
Check for existing locks:
rc lock list
Tips
- Use
rc context --compact for a token-efficient summary
- Use
rc search "keyword" to find relevant epics, issues, or wiki pages
- Always read comments before starting — the user may have specific instructions
- Update the wiki with architectural decisions or important context for future agents