Skip to content

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

IntegrationWhat GESA LearnsWhat GESA Influences
GESA × DRIFTGap trajectory across episodesStrategy risk tolerance
GESA × FetchWhich decision thresholds produce good outcomesThreshold calibration recommendations
GESA × 6D ForagingDimension-specific intervention effectivenessPer-dimension strategy ranking
GESA × HEATTeam-level effort patterns and pain signalsWorkplace intervention recommendations

Common Principle Across All Integrations

Every integration follows the same pattern:

  1. The upstream layer generates an episode input — a DRIFT score, a Fetch decision, a HEAT pain signal, a 6D cascade analysis
  2. GESA captures it as an episode — with full context, action, and pending outcome
  3. GESA's episode history informs future cycles — the same situation will trigger retrieval of this episode as a relevant reference
  4. 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.


→ GESA × DRIFT