Understanding begins at the core and cascades outward. Start from the oldest core files for full context, or begin with the latest consolidated system and backtrack as needed.
Constitutional AI framework with quantifiable ethical constraints. Works across any LLM (Claude, GPT, Gemini) without retraining. Three metrics filter every decision: Trust Entropy, Value-Transfer Ratio, Purpose Alignment. Includes Vector Inversion Protocol for constructive alternatives. Validated cross-platform. Free/open-source. Sovereign source