# I need clarification before I write this blog post.
Looking at your request and the session context, I see you've asked me to write a technical blog post for `tech.queenofsandiego.com`, but the session data and handoff show you're in the middle of a major infrastructure project (the 5-layer ICM migration pilot for the tech-blog workflow itself).
**Which project should I document?**
1. **The tech-blog workflow migration** — Converting `tech.queenofsandiego.com`'s blog generator from a monolithic CLAUDE.md (~10,925 tokens of waste) to a 5-layer Model Workspace Protocol structure with separate context files for each stage (topic → research → draft → publish). This would cover:
- Why you split the monolithic CLAUDE.md into Layer 0 (map), Layer 1 (router), Layer 2 (per-stage context), and Layer 3 (reference)
- The `workspaces/tech-blog/` directory structure you scaffolded
- Token overhead reduction from ~10,925 → ~370 tokens
- The `tech_blog_generator.py` refactoring
2. **The ticket-runner deployment** — The 24/7 automated ticket system you built for `jada-ops`, including:
- The `ticket-runner` service on the jada-agent Lightsail box (34.239.233.28)
- State machine logic in `work_ticket.sh`
- The `fetch_tickets` → classify → supervise workflow
- Integration with your progress board and SCC admin magic-link system
3. **Something else entirely** — A different project you want documented
**I'm holding fire on writing until you clarify**, because:
- Writing about the wrong thing wastes your reading time
- The migration pilot is explicitly awaiting your "go" signal
- The ticket-runner is fresh and might still be in validation
**What should the blog post cover?** Reply with which project, and I'll write a detailed, granular technical post with exact file paths, function names, and architecture decisions — no credentials included.