awesome-repositories.com
Blog
MCP
awesome-repositories.com

Discover the best open-source repositories with AI-powered search.

ExploreCurated searchesOpen-source alternativesSelf-hosted softwareBlogSitemap
ProjectMCP serverAboutHow we rankPress
LegalPrivacyTerms
© 2026 Bringes Technology SRL·VAT RO45896025·hello@awesome-repositories.com
microsoft avatar

microsoft/agent-governance-toolkit

0
View on GitHub↗

Agent Governance Toolkit

The agent-governance-toolkit is a framework for enforcing security policies, managing zero-trust identities, and sandboxing the execution of autonomous AI agents. It provides a governance layer designed to control the behavior of agents through the use of a security policy engine, cryptographic identity management, and a runtime execution sandbox.

The project distinguishes itself through a multi-tier privilege ring system and a cryptographic identity mesh that secures communication between autonomous entities. It implements a decay-based trust scoring mechanism to track entity reliability and utilizes hash-chained, tamper-proof audit logs to maintain a verifiable history of execution.

The toolkit covers a broad range of capability areas, including prompt security to defend against injection attacks, automated compliance mapping against regulatory standards, and autonomous workflow orchestration using saga patterns. It also features fleet monitoring for tracking health and spending limits, as well as tool execution sandboxing to restrict unauthorized resource access.

A command-line interface is provided for executing control signals, validating governance policies, and managing the installation of extensions.

AI search

Explore more awesome repositories

Describe what you need in plain English — the AI ranks thousands of curated open-source projects by relevance.

Start searching with AI

Features

  • Agent Governance Frameworks - Provides a comprehensive framework for enforcing security policies, managing zero-trust identities, and sandboxing autonomous AI agents.
  • AI Governance Tools - Implements a governance layer that intercepts AI tool calls to enforce security policies and operational rules.
  • Agent Governance - Establishes security policies and guardrails to control the behavior and execution of autonomous AI agents.
  • Agent Discovery - Provides capabilities for searching and locating specific agent instances within a secure trust mesh.
  • AI Execution Sandboxes - Isolates tool calls and system commands through a privilege ring system to prevent unauthorized resource access.
  • Autonomous Agent Orchestration - Implements a system for managing complex multi-step agent transactions using saga patterns to ensure state consistency.
  • Process Kill Switches - Provides emergency mechanisms for immediate process termination of autonomous agents to mitigate risks.
  • Action Policy Enforcement - Intercepts agent tool calls to enforce governance rules, including mandatory human-in-the-loop approvals.
  • Agent Communication Protocols - Provides a mechanism to unify trust models across communication standards for consistent identity and security.
  • Agent Execution Terminations - Provides an immediate kill switch to stop autonomous entities while maintaining a detailed audit trail.
  • LLM Tool Proxies - Proxies requests to tool servers to apply security policy checks and sanitize outputs before they reach the LLM.
  • Execution Privilege Rings - Assigns agents to hierarchical execution levels to restrict access to sensitive system resources.
  • Execution Sandboxes - Implements isolated execution environments that restrict system calls and file access using privilege rings to prevent unauthorized operations.
  • Prompt Injection Detectors - Uses a multi-vector evaluation system to detect and block prompt injection and jailbreak attempts.
  • Agent Execution Kill Switches - Provides a scoped containment mechanism to halt agent execution across global, per-entity, or per-capability levels.
  • Agent Security - Identifies and mitigates vulnerabilities in AI agents based on industry standards like the OWASP Agentic Top 10.
  • Agentic Risk Mitigations - Provides a comprehensive protection suite combining policy engines, sandboxing, and execution rings to mitigate agentic risks.
  • Agentic Security Standard Verifications - Scans agent configurations and prompts to identify vulnerabilities and validate adherence to OWASP Agentic Top 10 standards.
  • Decay-Based Trust Scores - Tracks entity reliability using a time-decaying scoring system based on successful operations and policy violations.
  • Cryptographic Identity Generation - Assigns unique identifiers and decentralized IDs to entities to ensure accountability through human sponsorship.
  • Data Redaction Tools - Provides a scanner that removes secret patterns from tool outputs before they are processed by the AI model.
  • AI Governance Policies - Implements a declarative security policy engine that intercepts tool calls and prompts to ensure OWASP compliance.
  • Framework Wrappers - Provides a governance layer that wraps existing AI frameworks to apply policy enforcement and security controls.
  • Agent Identities - Manages unique identities and trust mesh routing to ensure only verified agents interact.
  • Cryptographic Identity Networks - Establishes a trust mesh using handshakes and decentralized identifiers to secure inter-agent communication.
  • Security Policy Enforcers - Implements an immutable policy engine to prevent goal hijacking and unauthorized code execution.
  • Agent Execution Profiles - Restricts risky shell operations and credential access using tiered security profiles for agent execution.
  • Peer Trust Scoring - Implements a decay-based trust scoring system to track entity reliability based on task success and policy compliance.
  • LLM Prompt Injection Prevention - Scans prompts for injection and poisoning patterns to prevent adversarial attacks on the agent.
  • Policy-Based Access Control - Utilizes a rule-based engine to grant or deny access to resources based on defined authorization policies.
  • Agent Privilege Models - Implements a multi-level privilege model that requires human approval for high-risk autonomous operations.
  • Prompt Risk Scoring - Scores the risk level of prompts by detecting adversarial injection and exfiltration patterns.
  • Threat-Based Scanning - Implements a malicious pattern detector that blocks the execution of dangerous user prompts.
  • Credential Redaction - Implements pattern-based masking to redact sensitive credentials from agent communications and logs.
  • Task-Specific Storage Isolators - Implements a virtual file system that provides isolated memory and working contexts for individual AI entities.
  • Trust Verification - Implements a handshake mechanism to validate identities and trust scores before allowing agents to communicate.
  • Agent Storage Isolation - Provides isolated virtual file systems to separate memory and working contexts for individual agents.
  • Request Interception Middleware - Implements a composable middleware stack to wrap operations for security policy enforcement and auditing.
  • Composable Middleware Pipelines - Utilizes configurable chains of reusable functions to enforce security policies and audit incoming requests.
  • Agent Fleet Management - Tracks the health, registry, and heartbeats of multiple AI entities to monitor aggregate system status.
  • Agent Fleet Monitoring Dashboards - Provides a real-time visibility dashboard for the health, trust levels, and compliance status of active agent fleets.
  • Shadow Agent Detection - Scans processes, configurations, and repositories to identify unregistered AI entities within the system.
  • Action Auditing - Provides detailed records of agent interactions and tool executions to maintain a verifiable history of activity.
  • Definition Security Scanners - Provides an inspection utility that checks tool definitions for poisoning, typosquatting, and hidden instructions.
  • API Operational Cost Limits - Enforces hard spending limits and applies throttling or kill-switches to prevent AI operational budget overruns.
  • Agent Reliability Trackers - Tracks agent success rates, tool accuracy, and hallucination rates against defined reliability targets.
  • Agent Memory Management - Provides a storage system for retrieving episodic conversation history to maintain context across interaction turns.
  • Collaborative Session Coordination - Implements a coordination system for shared execution environments with configurable consistency modes for state synchronization.
  • Sandbox Graceful Shutdowns - Implements a trigger-based system for stopping execution due to drift or timeouts while handling pending transactions.
  • Multi-Protocol Transaction Orchestration - Provides an orchestration system using the Saga pattern to coordinate distributed transactions across complex action sequences.
  • Model Tool Schemas - Converts tool definitions into model-specific formats for large language model understanding and execution.
  • Compliance Mapping Tools - Automatically maps agent actions against regulatory frameworks and generates adherence reports.
  • Compliance Verification Tools - Provides an integrity check and policy linting system to ensure agent adherence to security standards.
  • Credential Lifecycle Management - Manages the provisioning, rotation, and revocation of ephemeral credentials using quantum-safe signatures.
  • Block Chaining - Records activities in an append-only SHA-256 hash chain to create a tamper-proof execution history.
  • Command Restrictions - Validates subprocess commands against a denylist and strips shell metacharacters to prevent destructive operations.
  • Regulatory Compliance - Provides systems for evaluating agent adherence to legal and regulatory frameworks.
  • Action Risk Classifications - Assigns risk levels to automated agent actions based on their potential impact on the environment and regulatory domains.
  • Multi-Party Authorizations - Implements a quorum-based approval system for high-risk operations to prevent single points of failure and approval fatigue.
  • Immutable Policy Locking - Freezes the security policy engine after initialization to prevent unauthorized runtime modifications.
  • Joint Liability Systems - Implements a reputation system that links sponsors to sponsored entities to penalize both during governance violations.
  • LLM Governance and Compliance - Implements a compliance grading system to measure the effectiveness of active security policies and risk mitigations.
  • Hash-Chained Event Logs - Ships a SHA-256 hash-chained audit log system used to detect unauthorized modifications to execution history.
  • System Prompt Validations - Provides a deterministic pattern scanner to verify that defensive language is present in system prompts before deployment.
  • Policy Validators - Provides schema compliance checks for governance policy files and plugin manifests to ensure correctness.
  • Ephemeral Privilege Elevations - Grants short-term increases in access levels that automatically revert after a set period or task completion.
  • Integrity Verifications - Uses a registration-time hashing system to verify tool integrity and block tampered calls before execution.
  • Governance Evaluations - Verifies plugin manifests against governance files to ensure adherence to security and operational criteria.
  • Saga Coordinators - Manages long-running agent workflows using saga patterns with automatic compensation logic for consistency.
  • Integrity Verifications - Implements cryptographic checksums and key verification to ensure plugin manifests remain unaltered before execution.
  • State Consistency Controllers - Ensures shared state consistency through the use of version counters, resource locks, and isolation levels.
  • Collaborative Liability Tracking - Implements a provenance system that assigns joint liability across collaborative interactions to determine entity responsibility.
  • Agent Health Monitoring - Provides tools for tracking the status and connectivity of AI agents with integrated failure kill switches.
  • Agent Fleet Telemetry - Provides a dashboard and telemetry system for tracking agent health, spending limits, and reliability metrics across fleets.
  • Agent Performance Monitoring - Implements a monitoring suite to track operational metrics, system health, and behavior of autonomous agents.
  • AI Cost Monitoring - Provides a statistical system for detecting spending anomalies and enforcing hierarchical budgets for AI operations.
  • System Quality Evaluators - Provides a scoring system that measures agent performance against reliability tiers using judge-based evaluations.
  • Tamper-Proof - Creates hash-chained, append-only logs of activities to ensure a verifiable and tamper-proof record of execution.
  • Service Level Objective Management - Provides utilities to compute attainment and error budgets by tracking request outcomes and latency for agent services.
  • AI Agent Behavior Monitors - Tracks the health, cost, and performance of agent fleets using real-time telemetry and emergency kill switches.
  • Security and Red Teaming - Policy enforcement and zero-trust identity for agentic systems.
6,535 stars·1,129 forks·Python·MIT·47 views

Star history

Star history chart for microsoft/agent-governance-toolkitStar history chart for microsoft/agent-governance-toolkit

How this analysis was created: This summary and feature list are AI-generated from collected project material and can contain mistakes. Stars, license and language are imported from GitHub. Inclusion does not mean that we have tested or audited this project. Check the source documentation for any feature you depend on. Learn more on our About page.

Frequently asked questions

What does microsoft/agent-governance-toolkit do?

The agent-governance-toolkit is a framework for enforcing security policies, managing zero-trust identities, and sandboxing the execution of autonomous AI agents. It provides a governance layer designed to control the behavior of agents through the use of a security policy engine, cryptographic identity management, and a runtime execution sandbox.

What are the main features of microsoft/agent-governance-toolkit?

The main features of microsoft/agent-governance-toolkit are: Agent Governance Frameworks, AI Governance Tools, Agent Governance, Agent Discovery, AI Execution Sandboxes, Autonomous Agent Orchestration, Process Kill Switches, Action Policy Enforcement.

Which projects share features with microsoft/agent-governance-toolkit?

Projects with overlapping indexed features include: atmosphere/atmosphere — Atmosphere is a Java-based framework for building and coordinating AI agents. It provides a real-time transport layer… mastra-ai/mastra — Mastra is an orchestration framework designed for building, deploying, and managing autonomous AI agents and… fetchai/innovation-lab-examples — This project provides a comprehensive framework for building, deploying, and orchestrating autonomous agents within a… github/awesome-copilot — Awesome Copilot is a comprehensive framework for autonomous software development, providing the infrastructure to… kortix-ai/suna — Suna is an orchestration platform designed for the deployment, management, and governance of autonomous AI agents. It… build-trust/ockam — Ockam is a zero-trust networking framework designed to secure data transit between distributed applications using an…

Projects sharing features with Agent Governance Toolkit

These projects share indexed features with Agent Governance Toolkit. Shared tags can include platform or build tooling; verify the primary use case before treating a result as a replacement.
  • atmosphere/atmosphereAtmosphere avatar

    Atmosphere/atmosphere

    3,830View on GitHub↗

    Atmosphere is a Java-based framework for building and coordinating AI agents. It provides a real-time transport layer for streaming data via WebSockets, SSE, gRPC, and WebTransport, alongside a multi-agent orchestration framework for managing agent fleets through sequential, parallel, and graph-based execution workflows. The project features a durable workflow engine that persists agent state as snapshots, allowing long-running tasks to survive system restarts and incorporate human-in-the-loop approvals. It also implements Model Context Protocol servers to expose tools, resources, and prompt

    Javaacpagentic-aiembabel
    View on GitHub↗3,830
  • fetchai/innovation-lab-examplesfetchai avatar

    fetchai/innovation-lab-examples

    1,028View on GitHub↗

    This project provides a comprehensive framework for building, deploying, and orchestrating autonomous agents within a decentralized network. It serves as a collection of patterns and examples for developing intelligent software entities capable of performing complex tasks, making decisions, and interacting with other agents to achieve shared goals. The framework distinguishes itself through its focus on multi-agent orchestration and decentralized communication. It enables the coordination of specialized agent teams that collaborate on workflows through structured messaging protocols, allowing

    Python
    View on GitHub↗1,028
  • mastra-ai/mastramastra-ai avatar

    mastra-ai/mastra

    21,221View on GitHub↗

    Mastra is an orchestration framework designed for building, deploying, and managing autonomous AI agents and multi-agent systems. It provides a comprehensive suite of primitives for creating resilient AI applications, including durable workflow orchestration, event-driven agent loops, and semantic memory management. By integrating these core components, the platform enables developers to build complex, multi-step processes that can reason about goals and execute tasks without manual intervention. The framework distinguishes itself through its focus on observability and secure, isolated execut

    TypeScriptagentsaichatbots
    View on GitHub↗21,221
  • github/awesome-copilotgithub avatar

    github/awesome-copilot

    40,904View on GitHub↗

    Awesome Copilot is a comprehensive framework for autonomous software development, providing the infrastructure to orchestrate multi-agent teams and automate complex coding workflows. It functions as a centralized platform for managing AI-driven development, enabling developers to deploy specialized agents that interact with local files, terminal commands, and external APIs to execute end-to-end software delivery tasks. The project distinguishes itself through its focus on governance and extensibility, offering a suite of security controls, policy-based execution guardrails, and audit trails t

    JavaScriptaigithub-copilothacktoberfest
    View on GitHub↗40,904
Compare all 30 related projects→