Parallel Implementation Tracking

Use this recipe when multiple agents implement a roadmap in parallel and leadership needs continuous visibility.

Setup

  • Project: Priivacy Rust recognizers
  • Active worktree: .worktrees/001-systematic-recognizer-enhancement
  • Dashboard URL: http://localhost:3000 (or custom port from spec-kitty dashboard)

Steps

  1. Start Dashboard – Run spec-kitty dashboard to launch the real-time kanban view. Dashboard runs in background and auto-refreshes.

  2. Snapshot lane counts – Dashboard shows items in planned, doing, for_review, done with live updates. Take screenshots for hourly reports.

  3. Move prompts via workflow commands – Always use spec-kitty agent action implement/review so the dashboard stays synchronized:

    spec-kitty agent action implement WP01

  4. Record activity logs – Agents append ISO 8601 entries to the prompt's "Activity Log" section for auditability:

    ## Activity Log
- 2025-01-15T09:30:00Z – claude – shell_pid=12345 – lane=doing – Started implementation
- 2025-01-15T11:45:00Z – claude – shell_pid=12345 – lane=for_review – Ready for review

  5. Monitor task completion – Review kitty-specs/<feature>/tasks.md checklist to ensure all subtasks are checked before merge.

  6. Automate alerts (Optional) – Use dashboard API endpoints for monitoring:

    • GET /api/features - List all features and their work packages
    • GET /api/feature/{slug} - Get specific feature details
    • Build custom alerts when tasks spend >4 hours in doing lane

Reporting

  • Export tasks.md and dashboard screenshots at daily stand-up
  • Summarize agent throughput using the Activity Log entries in work package files
  • Identify bottlenecks by checking lane distribution in dashboard
  • Use /spec-kitty.accept --mode checklist to generate readiness report
  • Use /spec-kitty.merge --dry-run to produce merge preview for executives

Dashboard Features for Tracking

  • Real-time updates - Refresh is automatic via polling, no manual page reload needed
  • Lane filtering - Focus on specific lanes (e.g., only "for_review")
  • Agent assignments - See which agent is working on which task
  • Completion metrics - Track progress percentages per feature
Edit this page