Memory Save Skill

Save current sprint context before ending a session, so work can be resumed later.

Instructions

1. •

   Get current context:

   • •Run git branch --show-current to get branch name
   • •Extract sprint number from branch (e.g., "Sprint_12" → "Sprint 12")
   • •Get current date/time
2. •

   Gather session context from the conversation:

   • •Current tasks (from todo list or discussion)
   • •Recent work completed this session
   • •Next steps to continue
   • •Any blockers or important notes
3. •

   Write the memory file to .claude/memory/current.md with this structure:

   markdown

   # Sprint Context Save
   
   **Sprint**: [sprint name]
   **Date**: [YYYY-MM-DD HH:MM:SS]
   **Branch**: [branch name]
   **Status**: In Progress
   
   ## Current Tasks
   
   - [ ] Task 1
   - [x] Completed task
   
   ## Recent Work
   
   - What was done this session
   
   ## Next Steps
   
   - What to do when resuming
   
   ## Blockers/Notes
   
   Any blockers or important context
   
   ---
   
   **Instructions for Claude on Resume**:
   1. Read this context file on startup
   2. Verify git branch matches sprint
   3. Continue from "Next Steps" section above
   

4. •

   Update metadata at .claude/memory/memory_metadata.json:

   json

   {
     "current_save": ".claude/memory/current.md",
     "last_updated": "[ISO timestamp]",
     "sprint": "[sprint name]",
     "status": "active",
     "pending_restore": true
   }
   

   IMPORTANT: The pending_restore: true flag signals that this save should be restored on next startup. The /startup-check skill will clear this flag after restoring.

5. •

   Confirm save by showing the user a summary

Output Format

Memory Saved:
- Sprint: [sprint name]
- Branch: [branch name]
- Tasks: [N] tasks recorded
- Next Steps: [brief summary]
- File: .claude/memory/current.md

Ready to resume with /memory-restore in a new session.