A runtime and builder enabling direct Python execution inside GGUF-based LLMs using the 4DLLM container format. Supports modular cognition, dependency-aware module loading, and programmable model behavior at inference time.