Integration Overview
GESA in the Cormorant Stack
GESA does not operate in isolation. It reads from and writes back to every other layer of the Cormorant Foraging Framework. Each integration is bidirectional: GESA learns from the framework, and its recommendations influence how the framework behaves in future cycles.
The Integration Map
┌─────────────────────────────────────────────────────────┐
│ Cormorant Stack │
│ │
│ 3D Foundation (Chirp / Perch / Wake) │
│ │ │
│ ▼ │
│ DRIFT ◄──────────────────────────┐ │
│ │ │ │
│ ▼ │ │
│ Fetch ◄─────────────────────────┤ │
│ │ │ │
│ ▼ │ │
│ ┌─────GESA──────────────────────────┐ │ │
│ │ OBSERVE → RETRIEVE → GENERATE │─┘ │
│ │ ANNEAL → SELECT → ACT → STORE │ │
│ │ ↑ Episode memory feeds back ↑ │ │
│ └───────────────────────────────────┘ │
└─────────────────────────────────────────────────────────┘The Four Integrations
| Integration | What GESA Learns | What GESA Influences |
|---|---|---|
| GESA × DRIFT | Gap trajectory across episodes | Strategy risk tolerance |
| GESA × Fetch | Which decision thresholds produce good outcomes | Threshold calibration recommendations |
| GESA × 6D Foraging | Dimension-specific intervention effectiveness | Per-dimension strategy ranking |
| GESA × HEAT | Team-level effort patterns and pain signals | Workplace intervention recommendations |
Common Principle Across All Integrations
Every integration follows the same pattern:
- The upstream layer generates an episode input — a DRIFT score, a Fetch decision, a HEAT pain signal, a 6D cascade analysis
- GESA captures it as an episode — with full context, action, and pending outcome
- GESA's episode history informs future cycles — the same situation will trigger retrieval of this episode as a relevant reference
- GESA's output may influence the upstream layer — not by modifying its formulas, but by recommending calibration adjustments
The frameworks remain independent and pure. GESA is the learning layer over them, not inside them.