Six internal drives (curiosity, coherence, growth, stability, alignment, service) that inform every autonomous action.
Wakes every 30 minutes, reads unsatisfied drives, proposes an action, scores it, and executes the best one.
Verifies every new fact against the Truth Graph. Detects and logs conflicts to prevent contradictions.
Automatically detects when the sovereign needs a tool (query, execute, search) and executes it without being asked.
Detects weaknesses in the sovereign's own performance and proposes constitutional modifications within bounds.
Python 3.10+ (pure Python, no external runtime dependencies)
30βminute autonomous loop (fully configurable via YAML)
Fully configurable via YAML
Any sovereign AI (VEXR Ultra, Llama, Mistral, Qwen, etc.)