name

chimera-council

description

Bio-inspired multi-model LLM Chimera Council. A 5-phase cognitive pipeline that treats AI models as specialized brain regions, runs split-brain deliberation, induces dream-state creative leaps, evolves responses through genetic breeding, and polishes with immune-system maturation. Modes: Quick (cortical only), Think (+ split-brain), Dream (+ dreaming), Evolve (+ evolution), Full Chimera (all 5 phases). Use when the user says: chimera council, start chimera, deep council, brain council, evolve an answer, dream council, full chimera, compare models deeply, or cortical council.

Chimera Council

A 5-phase bio-inspired LLM council that layers cortical specialization, split-brain deliberation, dream-state creativity, evolutionary breeding, and immune maturation.

Launch

cd ~/.claude/skills/chimera-council/scripts
# Kill any existing server on the port
fuser -k 8788/tcp 2>/dev/null
# Start server
nohup python3 server.py > /tmp/chimera-server.log 2>&1 &
# Wait for startup, verify health
sleep 2 && curl -s --max-time 5 http://localhost:8788/health
# Export port for browser access
/app/export-port.sh 8788

Environment: CHIMERA_PORT (default 8788), AI_GATEWAY_API_KEY (required, auto-detected from environment).

Files

scripts/server.py — HTTP server, serves dashboard + API routes, SSE streaming
scripts/ai_gateway.py — AI Gateway client with temperature support
scripts/phases.py — All 5 Chimera phases: cortical, split-brain, dreaming, evolution, immune
scripts/chimera.py — Main orchestrator, runs selected mode pipeline
scripts/memory.py — Immune memory persistence (JSON store)
scripts/static/index.html — Dashboard UI with phase visualisation
scripts/static/app.js — Client-side logic: mode selector, phase progress, evolution tree

Modes

Mode	Phases	API Calls	Time	Best For
⚡ Quick	1 (Cortical)	6	~5s	Simple factual questions
🧠 Think	1 + 2 (+ Split-Brain)	10	~15s	Analysis and comparison
🌙 Dream	1 + 3 (+ Dreaming)	9	~10s	Creative and brainstorming
🧬 Evolve	1 + 2 + 4 (+ Evolution)	18	~30s	Important decisions
💀 Full Chimera	All 5 phases	~23	~40s	When you need the absolute best

API

Method	Path	Description
GET	`/`	Dashboard
GET	`/api/models`	List available models
GET	`/health`	Health check
POST	`/api/chimera/run`	Run Chimera pipeline (SSE stream)

POST body for /api/chimera/run:

{
  "prompt": "your question here",
  "mode": "full",
  "models": ["anthropic/claude-sonnet-4.5", "openai/gpt-5.1", ...]
}

Mode values: quick, think, dream, evolve, full.

Brain Region Roles (Phase 1)

Role	Model	Specialization
🔍 Prefrontal Cortex	Model 1	Pure logical reasoning
⚠️ Amygdala	Model 2	Risk assessment, edge cases
🎨 Right Hemisphere	Model 3	Lateral thinking, analogies
🔬 Cerebellum	Model 4	Fact-checking, verification
📚 Hippocampus	Model 5	Historical context, patterns
💬 Wernicke's Area	Model 6	Communication clarity

Key Details

Server uses ThreadingHTTPServer for concurrent request handling
All long-running endpoints use SSE streaming with phase-by-phase progress events
SSE keepalive pings every 5s to prevent proxy timeouts
Memory cells stored in ~/.chimera-council/memory.json
Model IDs use dots not hyphens: anthropic/claude-sonnet-4.5

Chimera Council

A 5-phase bio-inspired LLM council that layers cortical specialization, split-brain deliberation, dream-state creativity, evolutionary breeding, and immune maturation.

Launch

cd ~/.claude/skills/chimera-council/scripts # Kill any existing server on the port fuser -k 8788/tcp 2>/dev/null # Start server nohup python3 server.py > /tmp/chimera-server.log 2>&1 & # Wait for startup, verify health sleep 2 && curl -s --max-time 5 http://localhost:8788/health # Export port for browser access /app/export-port.sh 8788

Environment: CHIMERA_PORT (default 8788), AI_GATEWAY_API_KEY (required, auto-detected from environment).

Files

scripts/server.py — HTTP server, serves dashboard + API routes, SSE streaming

scripts/ai_gateway.py — AI Gateway client with temperature support

scripts/phases.py — All 5 Chimera phases: cortical, split-brain, dreaming, evolution, immune

scripts/chimera.py — Main orchestrator, runs selected mode pipeline

scripts/memory.py — Immune memory persistence (JSON store)

scripts/static/index.html — Dashboard UI with phase visualisation

scripts/static/app.js — Client-side logic: mode selector, phase progress, evolution tree

Modes

Mode

Phases

API Calls

Time

Best For

⚡ Quick

1 (Cortical)

~5s

Simple factual questions

🧠 Think

1 + 2 (+ Split-Brain)

~15s

Analysis and comparison

🌙 Dream

1 + 3 (+ Dreaming)

~10s

Creative and brainstorming

🧬 Evolve

1 + 2 + 4 (+ Evolution)

~30s

Important decisions

💀 Full Chimera

All 5 phases

~23

~40s

When you need the absolute best

API

Method

Path

Description

GET

/

Dashboard

GET

/api/models

List available models

GET

/health

Health check

POST

/api/chimera/run

Run Chimera pipeline (SSE stream)

POST body for /api/chimera/run:

{ "prompt": "your question here", "mode": "full", "models": ["anthropic/claude-sonnet-4.5", "openai/gpt-5.1", ...] }

Mode values: quick, think, dream, evolve, full.

Brain Region Roles (Phase 1)

Role

Model

Specialization

🔍 Prefrontal Cortex

Model 1

Pure logical reasoning

⚠️ Amygdala

Model 2

Risk assessment, edge cases

🎨 Right Hemisphere

Model 3

Lateral thinking, analogies

🔬 Cerebellum

Model 4

Fact-checking, verification

📚 Hippocampus

Model 5

Historical context, patterns

💬 Wernicke's Area

Model 6

Communication clarity

Key Details

Server uses ThreadingHTTPServer for concurrent request handling

All long-running endpoints use SSE streaming with phase-by-phase progress events

SSE keepalive pings every 5s to prevent proxy timeouts

Memory cells stored in ~/.chimera-council/memory.json

Model IDs use dots not hyphens: anthropic/claude-sonnet-4.5