microsoft/agent-governance-toolkit
AI Agent Governance Toolkit — Policy enforcement, zero-trust identity, execution sandboxing, and reliability engineering for autonomous AI agents. Covers 10/10 OWASP Agentic Top 10.
What it solves
It addresses the risk of autonomous AI agents performing unauthorized or destructive actions. Because prompt-level safety is probabilistic and susceptible to injection attacks, this toolkit provides a deterministic layer of enforcement that intercepts tool calls and messages before they are executed, making unauthorized actions structurally impossible rather than just unlikely.
How it works
The toolkit acts as a governance kernel between the AI agent and its tools. It uses a policy engine (supporting YAML, OPA, and Cedar) to evaluate every requested action against a set of rules. If an action is allowed, it proceeds; if denied, it raises a GovernanceDenied error. The system also integrates identity management (SPIFFE/DID) and tamper-evident audit logs to track exactly which agent performed which action and why it was permitted.
Who it’s for
Developers and SREs shipping autonomous agents to production who need strict policy enforcement, regulatory compliance (such as OWASP Agentic Top 10), and detailed auditability across multi-agent systems.
Highlights
- Multi-Language Support: SDKs available for Python, TypeScript, .NET, Rust, and Go.
- Framework Agnostic: Integrates with major frameworks including Semantic Kernel, AutoGen, LangGraph, CrewAI, and OpenAI Agents SDK.
- Deterministic Enforcement: Moves safety from the prompt layer to the application code layer to prevent prompt injection bypasses.
- Comprehensive Tooling: Includes a CLI for OWASP compliance checks, policy linting, and prompt injection auditing.
- Advanced Governance: Features like execution sandboxing (privilege rings), a kill switch, and an MCP Security Gateway for tool poisoning detection.