GateGraph – deterministic governance for AI agents
GateGraph is a deterministic governance layer designed for AI-agent actions, focusing on evaluating requests before execution. It ensures that execution authority remains separate from the model while providing bounded governance decisions. The current release emphasizes local, single-node operation without public internet exposure or autonomous capabilities.
- ▪GateGraph evaluates action requests and produces governance decisions such as allow, block, or review.
- ▪It is not an autonomous agent or a policy-learning system and does not support multi-node consensus.
- ▪The current release is limited to local operation and does not include public deployment capabilities.
Opening excerpt (first ~120 words) tap to expand
Release: v0.17.1_STABLE GateGraph Current release: v0.17.1_STABLE Base: v0.17.0_STABLE Base stable: v0.17.0_STABLE Status: candidate Version: 0.17.1 Phase: Promotion Status Drift Guard Release focus: Promotion Status Drift Guard GateGraph is a deterministic governance layer for AI-agent actions. It evaluates requested actions before execution, produces bounded governance decisions, and keeps execution authority outside the model. What GateGraph is GateGraph is designed to sit between an agent or operator workflow and an action surface. It can: evaluate action requests, return allow/block/review/approval decisions, issue bounded capability-token surfaces, produce audit and explainability artifacts, run deterministic evidence checks.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at GitHub.