Playbook / Principal search path / Spine health checklist

Spine health checklist

Rule I actually follow: warm the spine 15–20 minutes before every live technical call. Render Free cold starts (15–40s) are fine if I label them. Silent failure is not.

This stays true in chapter 4. I do not skip warmup because “I know these URLs.”


Warm these (in order)

Portfolio hub

Five-spine UIs

API /health (expect cold start)


What I say if something is cold

These APIs run on Render Free interim. First hit after idle can take 15–40 seconds. I’ll warm health while we talk architecture. Demo vs Strict is labeled on the UIs. For AuthZ and multi-tenant I walk the Strict/JWT design and the ADRs.

If health is 500, not just slow: don’t click through a broken demo. Walk the architecture and the Strict path. Patch later. Silent “it’s loading…” until they lose patience is worse than naming it.

If they ask to screen-share: two UIs, not five. Governance (AegisAI) or access-aware RAG — whichever matches their problem. Hub + technical-review if they want the map. I do not click the catalog.

If they say “just send me links”: hire + technical-review. That’s the signature. Not 17 product URLs.

If a UI 404s or auth loops mid-call:

I’m going to stop clicking and walk the architecture — gateway or authZ-before-rank, Demo vs Strict. I’ll patch after. Wrestling a dead tab for five minutes is worse than naming it.


Demo vs Strict — the lines I use

TopicI say
Enterprise RAG default liveOften Demo Principal — labeled
Production review pathStrict/JWT Principal; AuthZ before ranking
AegisAI registryIn-memory on free tier; Postgres path documented
Publish / side effectsFail-closed when PRODUCTION_STRICT

Outreach signature (only these)

Hire: https://venkat-ai.com/hire
Technical review: https://venkat-ai.com/technical-review

I do not paste 17 product URLs. Not in chapter 1. Not in chapter 4.


Last verified

DateResult
2026-07-31Hub + 4 spine UIs + AegisAI/VAP/ERAG health: 200. ACF health timed out once then 200 after warm (~1s).