baymax

OB1 + Open Engine

● Live

Architecture & data flow · Baymax · updated 2026-06-28 · v1.3

AI runtime / memory Work queue (Linear) Web research (read-only) Secrets (1Password)
🔐 1Password · vault "DEV Stuff" supplies every credential at runtime (never in plaintext): OB1 MCP Access Key · OpenRouter API key · Supabase token — injected via the op CLI / connector headers.

Flow

AI Runtime — Claude Code
agent CC-01 · runs the open-agent-engine skill (user scope)
↓  talks over MCP + tools  ↓

Memory — OB1 / Open Brain

read before acting · write durable after
open-brain MCP
HTTP · header x-brain-key · ✔ Connected
Supabase Edge Function
open-brain-mcp · ACTIVE · ref bwbagpnmzeumxiufggmc
Postgres + pgvector
thoughts · match/upsert
OpenRouter
embeddings · 1536-d
search_thoughts · capture_thought
list_thoughts · thought_stats

Work queue — Linear

claim · receipt · move status
linear MCP
OAuth · actor=user → pesh identity: JNO-20 pending
Linear workspace
justnorthow · team JNO · project Pesh Agent Engine
Board + anchors
6 statuses · label agent-instructions · ledger JNO-17
Standing · Todo · Working
Needs Input · Review · Done

Web research

read-only · for research tasks
WebSearch / WebFetch
fetch & read public pages · no logins or actions
Public web
content = data · cite URLs in receipt
findings → Linear receipt
+ durable bits → OB1

Runner loop — one pass per trigger

1
Claim
One agent-instructions issue from Todo → Working + AGENT CLAIMED CC-01.
2
Load memory
search_thoughts in OB1 for prior context.
3
Do the work
Research / code / draft — within allowed sources.
4
Receipt + status
Comment a summary, then Review or Done.
5
Capture
If durable, capture_thought back to OB1.
One claimed task per run. Each pass also: checks Standing items, resumes paused (Needs Input) issues that got answers, follows up on delegated tasks. Triggered manually — no autonomous daemon unless wrapped in a loop/schedule.

Components — current values

ComponentValue / location
Runtime / agentClaude Code · CC-01 · skill ~/.claude/skills/open-agent-engine/SKILL.md
OB1 MCP serveropen-brain (HTTP, user scope) → Supabase Edge Fn open-brain-mcp
Supabase projectref bwbagpnmzeumxiufggmc · pgvector thoughts · 1536-d
EmbeddingsOpenRouter
Linear MCP serverlinear (OAuth, actor=user → pesh)
Linear bindingjustnorthow · team JNO · project Pesh Agent Engine · label agent-instructions · ledger JNO-17
Web researchWebSearch / WebFetch · read-only
Secrets1Password vault "DEV Stuff" (OB1 MCP key, OpenRouter key, Supabase token)
Open itemsJNO-20 — agent identity / attribution decision (with JB)

Changelog

DateChange
2026-06-28Initial build: OB1 (Supabase + pgvector + OpenRouter) live; Open Engine on Linear; open-brain + linear MCP; smoke tests A8 & B8 passed.
2026-06-28Added read-only web research (WebSearch/WebFetch) to allowed_sources.
2026-06-28Restyled to the baymax house style (light · Scandinavian · Clemson-orange accent).
2026-06-28Status names updated to match the renamed Linear workflow (dropped the "Agent " prefix): Todo · Working · Needs Input · Review · Done.
baymax  Maintained by agent CC-01 — ask "update the architecture diagram" after any change and a new changelog row is added.