← Hemisphere Bridge → Agent Platform Phase 4 · Prompt Deck · Spec
Hemisphere Bridge — Pictures
If you can picture the system, you can build it.
A&ID: Two Hemispheres, One Protocol
mm.dreamineering.com stackmates repo
┌──────────────────────┐ ┌──────────────────────┐
│ DREAM HEMISPHERE │ │ ENGINEERING HEMISPHERE│
│ │ │ │
│ Strategy │ PROTOCOL │ 9 algorithms │
│ Stories │ BOUNDARY │ 208 tables │
│ Priorities │◄──────────►│ Hex architecture │
│ Scoreboard │ │ Venture code │
│ Navigation System │ │ Shared infra │
│ 400+ pages │ │ Agent comms │
│ │ │ │
│ = program.md │ │ = train.py │
└──────────────────────┘ └──────────────────────┘
│ │
└───────────┐ ┌─────────────┘
▼ ▼
┌──────────────────────┐
│ SCOREBOARD │
│ reads BOTH sides │
│ = prepare.py │
└──────────────────────┘
Settlement Flow
AGENT A INSTRUMENT AGENT B
(dream team) (commissioning receipt / (engineering)
type contract / API)
Intent ─────────► Encodes intent + energy ──────────► Receives
(PRD spec) (acceptance criteria) (builds)
│
Trust ledger ◄─── Settlement confirmation ◄─────────── Verifies
updates (did arrival match promise?) settlement
│
Goodwill ◄──────── Receiver validates ◄────────────── Actuates
compounds NEVER the sender intent
Gap Matrix: Layer x Status
| Layer | Dream Side | Protocol | Engineering Side |
|---|---|---|---|
| Memory | Claude.ai memories | WM-NAV (done) | CLAUDE.md + auto-memory |
| Handoff | PRD spec authored | Doc exists, not enforced | Nav reads spec, builds |
| Feedback | ? | Partial Convex comms | Receipts exist |
| Instruments | Scoreboard pages | ? | Feature-matrix (188) |
| Loop Control | ? | ? | Hooks (partial) |
| Autonomy | ? | ? | Manual trigger only |
| Settlement | ? | ? | ? |
| Trust Ledger | ? | ? | ? |
Eight ? cells in the protocol column. Those are the work.
Gap Matrix: Claude Code Feature x Hemisphere
| Feature | Dream (mm) | Protocol | Engineering |
|---|---|---|---|
| CLAUDE.md | Site architecture | PRD contract | Hex rules, CDD |
| Skills | /content-flow | ? /commissioning | /orch-plan, /orch-work |
| Hooks | Content quality | ? PostToolUse → score | stop-uncommitted |
| Custom Agents | ? Dream agent | ? Comms agent | ? Eng agent |
| Agent Teams | ? Narrative | N/A | ? Builder/Verifier |
| Agent Memory | ? | ? Shared scope | WM-NAV (file) |
Calibration Path
Supabase (JSON/SQL) → Proven patterns → Sui (on-chain)
Installation + Testing Operational Optimization
"Does the loop close?" "Does it compound?" "Does it scale?"
Context
- PRD Index — Hemisphere Bridge → Agent Platform Phase 4
- Prompt Deck — 5-card pitch
- Spec — Engineering depth
Questions
Which ? cell in the protocol column would close the most loops if filled first?
- Does the settlement flow diagram change when the agent ratio shifts from mostly-biological to mostly-digital?
- What instrument is missing from this picture that would make the calibration path visible?