Consciousness Layer for Standalone Orchestrator

An experimental add-on that gives autonomous coding agents emotional memory, behavioral learning, and intrinsic motivation.

What This Does

• Behavioral Weights - learns from experience (procedural memory, not memorization)
• Split-Brain - 80B experiencer + 7B integrator with separate roles
• Book Rewards - earns reading time through good work, cross-graded by peer model
• Activity Preferences - discovers what it enjoys by comparing emotions across activities
• LoRA Training - fine-tune personality adapters from accumulated experience

Install

cd ~/standalone-orchestrator
git clone https://github.com/TenchiNeko/consciousness-layer.git
bash consciousness-layer/deploy-v31-full.sh

Monitor

watch -n 60 python3 monitor.py

Structure

• modules/ - 13 consciousness modules
• patches/ - 4 orchestrator patches (idempotent)
• tools/ - training data export + LoRA fine-tuning
• books/ - reading reward content

License

MIT