When agents may act, and when they may not.

Six primitives governing autonomous execution: cryptographic policy enforcement, inference-time admissibility, confidence-governed suspension, capability-bounded action, curriculum-gated unlocking, and training-level governance. Non-execution is a first-class outcome.

Cryptographic Governance

Cryptographically signed policy agents, scoped mutation gating, and quorum-governed governance protocols. Compliance is structural.

51 articles

Inference Control

Semantic admissibility gate inside the inference loop. Every candidate transition evaluated against persistent agent state before commitment.

58 articles

Confidence Governance

Multi-input confidence computation. Rate of change detection. Non-executing cognitive mode. Hysteretic reauthorization.

59 articles

Capability Awareness

Capability envelope. Temporal executability forecasting. Joint condition: capability × time × uncertainty.

45 articles

LLM & Skill Gating

Unidirectional interface. Structural starvation. Curriculum-based progressive capability unlocking with evidence-based gates.

49 articles

Training Governance

Depth-selective gradient routing. Entropy-based profiles. Provenance-traceable training dynamics with zero-weight prevention.

42 articles