Skip to content
Back to search

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.

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today

Check it deeper on your machine

It has changed since its scan on 11 Nov 2026.

As a Claude Code plugin, install it and then ask Claude Code:

/plugin marketplace add keithmackay/mackayi/plugin install scanskillsafety@mackayiDo a deep scan of https://github.com/microsoft/agent-governance-toolkit with scanskillsafety

Or from a terminal, clone it and run it:

git clone https://github.com/keithmackay/scanskillsafety-skill ~/.claude/skills/scanskillsafetynode ~/.claude/skills/scanskillsafety/cli.cjs --deep 'https://github.com/microsoft/agent-governance-toolkit'

It checks the exact commit: every file, its packages against OSV (api.osv.dev), how it compares with this page's scan (findsafeskills.com receives the repository's address), and, if you agree, Claude reading every file with your own Claude subscription. It never runs the code.

Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/microsoft--agent-governance-toolkit/badge.svg)](/listings/microsoft--agent-governance-toolkit)

HTML

<a href="/listings/microsoft--agent-governance-toolkit"><img src="/api/listings/microsoft--agent-governance-toolkit/badge.svg" alt="findsafeskills safety rating" /></a>

Install

i

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

**Prerequisites:** Python 3.11+

pip install "agent-governance-toolkit[full]"

Use the `[full]` extra for the quick-start imports below. The base `agent-governance-toolkit` wheel installs the compliance CLI only; the governance modules live in the consolidated core distribution. The `agentmesh` quick-start import remains the current wrapper API. Importing `agent_os` emits a `DeprecationWarning` because the old `agent-os-kernel` distribution is deprecated. Use `agent-governance-toolkit-core` (or the `[full]` extra that includes it) as the replacement distribution. Policy-engine host code uses the ACS SDK; `agt-policies` provides the one-way v4-to-v5 migration command. The pre-ACS `agent_os.policies` rule model is gone, and `BREAKING_CHANGES.md` lists its replacements.

For Claude Code, add AGT as a plugin marketplace and install the governance plugin:

/plugin marketplace add microsoft/agent-governance-toolkit
/plugin install agt-governance@agent-governance-toolkit

Govern any tool function in two lines:

from agentmesh.governance import govern

safe_tool = govern(my_tool, policy="policy.yaml")   # every call checked, logged, enforced

On every call, `safe_tool` evaluates the YAML policy, logs the decision to an audit trail, and raises `GovernanceDenied` when the policy blocks the action.

# policy.yaml
apiVersion: governance.toolkit/v1
name: production-policy
default_action: allow
rules:
  - name: block-destructive
    condition: "action.type in ['drop', 'delete', 'truncate']"
    action: deny
    description: "Destructive operations require human approval"

  - name: require-approval-for-send
    condition: "action.type == 'send_email'"
    action: require_approval
    approvers: ["security-team"]
>>> safe_tool(action="read", table="users")
{'table': 'users', 'rows': 42}

>>> safe_tool(action="drop", table="users")
GovernanceDenied: Action denied by policy rule 'block-destructive':
  Destructive operations require human approval

Or use the full `AgentControl` API for programmatic control:

<details> <summary><b>AgentControl example</b></summary>

from agent_control_specification import AgentControl

runtime = AgentControl.from_path(str("manifest.yaml"))
result = runtime.evaluate(
    "input",
    {
        "envelope": {"agent_id": "example-agent"},
        "input": {"body": {"action": "web_search", "params": {}}},
    },
)
print(result.verdict)
runtime.close()

[Run the complete ACS email-tool example](examples/acs-email-tool).

</details>

<details> <summary><b>TypeScript / .NET / Rust / Go examples</b></summary>

**TypeScript**

import { PolicyEngine } from "@microsoft/agent-governance-sdk";

const engine = new PolicyEngine([
  { action: "web_search", effect: "allow" },
  { action: "shell_exec", effect: "deny" },
]);
engine.evaluate("web_search"); // "allow"
engine.evaluate("shell_exec"); // "deny"

**.NET**

using AgentGovernance;
using AgentGovernance.Extensions.ModelContextProtocol;
using AgentGovernance.Policy;

var kernel = new GovernanceKernel(new GovernanceOptions
{
    PolicyPaths = new() { "policies/default.yaml" },
});
var result = kernel.EvaluateToolCall("did:mesh:agent-1", "web_search",
    new() { ["query"] = "latest AI news" });

// MCP server integration
builder.Services.AddMcpServer()
    .WithGovernance(options => options.PolicyPaths.Add("policies/mcp.yaml"));

**Rust**

use agent_governance::{AgentMeshClient, ClientOptions};

let client = AgentMeshClient::new("my-agent").unwrap();
let result = client.execute_with_governance("data.read", None);
assert!(result.allowed);

**Go**

import agentmesh "github.com/microsoft/agent-governance-toolkit/agent-governance-golang"

client, _ := agentmesh.NewClient("my-agent",
    agentmesh.WithPolicyRules([]agentmesh.PolicyRule{
        {Action: "data.read", Effect: agentmesh.Allow},
        {Action: "*", Effect: agentmesh.Deny},
    }),
)

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

.claude-plugin/marketplace.json · Manifest · 32 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

This file has changed since it was scanned, so this is not exactly the text that was scanned, and no line is marked. The findings are listed with the safety scan above.

  1. {
  2. "name": "agent-governance-toolkit",
  3. "owner": {
  4. "name": "Microsoft Corporation",
  5. "email": "agentgovtoolkit@microsoft.com"
  6. },
  7. "description": "Agent Governance Toolkit plugins for Claude Code",
  8. "version": "5.0.0",
  9. "plugins": [
  10. {
  11. "name": "agt-governance",
  12. "source": "./agent-governance-claude-code",
  13. "description": "AGT governance hooks and MCP tools for Claude Code sessions",
  14. "version": "5.0.0",
  15. "author": {
  16. "name": "Microsoft Corporation",
  17. "email": "agentgovtoolkit@microsoft.com"
  18. },
  19. "homepage": "https://github.com/microsoft/agent-governance-toolkit/tree/main/agent-governance-claude-code",
  20. "repository": "https://github.com/microsoft/agent-governance-toolkit",
  21. "license": "MIT",
  22. "category": "security",
  23. "tags": [
  24. "agent-governance",
  25. "claude-code",
  26. "security",
  27. "policy",
  28. "mcp"
  29. ]
  30. }
  31. ]
  32. }
README.md · README · 469 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

This file has changed since it was scanned, so this is not exactly the text that was scanned, and no line is marked. The findings are listed with the safety scan above.

  1. 🌍 [English](/README.md) | [Español](./docs/i18n/README.es.md) | [日本語](./docs/i18n/README.ja.md) | [简体中文](./docs/i18n/README.zh-CN.md) | [한국어](./docs/i18n/README.ko.md) | [繁體中文](./docs/i18n/README.zh-TW.md)
  2. ![Agent Governance Toolkit](docs/assets/readme-banner.svg)
  3. # Agent Governance Toolkit
  4. ### Ship agents to production without losing sleep
  5. <p align="center">
  6. <a href="https://microsoft.github.io/agent-governance-toolkit">
  7. <img src="https://img.shields.io/badge/%F0%9F%93%96_Full_Documentation-microsoft.github.io%2Fagent--governance--toolkit-0078D4?style=for-the-badge&logoColor=white" alt="Full Documentation" height="40">
  8. </a>
  9. </p>
  10. <p align="center">
  11. <strong>
  12. 🚀 <a href="#quick-start">Quick Start</a> ·
  13. 📋 <a href="#specifications">Specifications</a> ·
  14. 📦 <a href="https://pypi.org/project/agent-governance-toolkit/">PyPI</a> ·
  15. 📝 <a href="CHANGELOG.md">Changelog</a>
  16. </strong>
  17. </p>
  18. [![CI](https://github.com/microsoft/agent-governance-toolkit/actions/workflows/ci.yml/badge.svg)](https://github.com/microsoft/agent-governance-toolkit/actions/workflows/ci.yml)
  19. [![Discord](https://dcbadge.limes.pink/api/server/TxMRqY3pFr?style=flat)](https://discord.gg/TxMRqY3pFr)
  20. [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](LICENSE)
  21. [![PyPI version](https://img.shields.io/pypi/v/agent-governance-toolkit?label=PyPI)](https://pypi.org/project/agent-governance-toolkit/)
  22. [![npm](https://img.shields.io/npm/v/%40microsoft/agent-governance-sdk?label=npm)](https://www.npmjs.com/package/@microsoft/agent-governance-sdk)
  23. [![NuGet](https://img.shields.io/nuget/v/Microsoft.AgentGovernance?label=NuGet)](https://www.nuget.org/packages/Microsoft.AgentGovernance)
  24. [![OpenSSF Scorecard](https://api.scorecard.dev/projects/github.com/microsoft/agent-governance-toolkit/badge)](https://scorecard.dev/viewer/?uri=github.com/microsoft/agent-governance-toolkit)
  25. [![OpenSSF Best Practices](https://www.bestpractices.dev/projects/12085/badge)](https://www.bestpractices.dev/projects/12085)
  26. [![OWASP Agentic Top 10](https://img.shields.io/badge/OWASP_Agentic_Top_10-7_Full,_3_Partial-blue)](docs/compliance/owasp-agentic-top10-architecture.md)
  27. [![AARM Extended](https://img.shields.io/badge/AARM-Extended_(R1–R9)-brightgreen)](https://aarm.dev/builders/agent-governance-toolkit-microsoft)
  28. [![ATF](https://img.shields.io/badge/ATF-All_5_Elements-brightgreen)](https://agentictrustframework.ai/ecosystem)
  29. Policy enforcement, identity, sandboxing, and SRE for autonomous AI agents. One `pip install`, any framework.
  30. ---
  31. ## The Problem
  32. Your AI agents call tools, browse the web, query databases, and delegate to other agents. Once deployed, they make decisions autonomously. You need answers to three questions:
  33. **1. Is this action allowed?** An agent with access to `send_email` and `query_database` should not be able to `drop_table`. OAuth scopes and IAM roles control which services an agent can reach, not what it does once connected.
  34. **2. Which agent did this?** In a multi-agent system, five agents might share a single API key. When something goes wrong, "an agent did it" is not an incident response.
  35. **3. Can you prove what happened?** Auditors and regulators need tamper-evident records of every decision: what policy was active, what the agent requested, and why it was allowed or denied.
  36. Prompt-level safety ("please follow the rules") is not a control surface. It is a polite request to a stochastic system. [OWASP LLM01:2025](https://genai.owasp.org/llmrisk/llm01-prompt-injection/) states this explicitly: *"it is unclear if there are fool-proof methods of prevention for prompt injection."* The published numbers back this up. [Andriushchenko et al. (ICLR 2025)](https://arxiv.org/abs/2404.02151) report **100% attack success rate** on GPT-4o, GPT-3.5, Claude 3, and Llama-3 using adaptive attacks with logprob access and suffix optimization, evaluated against the [JailbreakBench](https://arxiv.org/abs/2404.01318) benchmark (Chao et al., NeurIPS 2024). Microsoft's own [AI Red Teaming Agent](https://learn.microsoft.com/azure/ai-foundry/concepts/ai-red-teaming-agent) formalizes **Attack Success Rate (ASR)**, the rate of policy violations under adversarial input, as the canonical metric for this class of failure. [*Lessons from Red Teaming 100 Generative AI Products*](https://www.microsoft.com/en-us/security/blog/2025/01/13/3-takeaways-from-red-teaming-100-generative-ai-products/) reinforces the point: *"mitigations do not eliminate risk entirely"* and red teaming must be a continuous process because model-layer defenses are probabilistic by construction.
  37. AGT does not try to win that fight inside the prompt. Every tool call, message send, and delegation is intercepted in deterministic application code *before* the model's intent reaches the wire. Actions the AGT kernel denies are not "unlikely." They are **structurally impossible**. That is the difference between asking an agent to behave and making it incapable of misbehaving.
  38. ---
  39. ## Quick Start
  40. **Prerequisites:** Python 3.11+
  41. ```bash
  42. pip install "agent-governance-toolkit[full]"
  43. ```
  44. Use the `[full]` extra for the quick-start imports below. The base
  45. `agent-governance-toolkit` wheel installs the compliance CLI only; the governance
  46. modules live in the consolidated core distribution. The `agentmesh` quick-start
  47. import remains the current wrapper API. Importing `agent_os` emits a
  48. `DeprecationWarning` because the old `agent-os-kernel` distribution is deprecated.
  49. Use `agent-governance-toolkit-core` (or the `[full]` extra that includes it) as
  50. the replacement distribution. Policy-engine host code uses the ACS SDK;
  51. `agt-policies` provides the one-way v4-to-v5 migration command. The pre-ACS
  52. `agent_os.policies` rule model is gone, and `BREAKING_CHANGES.md` lists its
  53. replacements.
  54. The policy engine is published upstream as
  55. [`agent-control-spec`](https://github.com/responsibleai/agent-control-spec).
  56. [`agent-hooks`](https://github.com/responsibleai/agent-hooks) defines its
  57. interception contract. AGT retains the `AgentControl` host APIs and framework
  58. adapters in [`policy-engine/`](policy-engine/), not a second decision engine.
  59. This keeps policy evaluation in one place. AGT can adopt upstream fixes through
  60. a dependency update while keeping its host APIs and framework integrations.
  61. See the [ACS package guide](docs/packages/agent-control-specification.md) for
  62. the distinction between upstream packages and AGT's compatibility SDKs.
  63. For Claude Code, add AGT as a plugin marketplace and install the governance plugin:
  64. ```text
  65. /plugin marketplace add microsoft/agent-governance-toolkit
  66. /plugin install agt-governance@agent-governance-toolkit
  67. ```
  68. Govern any tool function in two lines:
  69. ```python
  70. from agentmesh.governance import govern
  71. safe_tool = govern(my_tool, policy="policy.yaml") # every call checked, logged, enforced
  72. ```
  73. On every call, `safe_tool` evaluates the YAML policy, logs the decision to an
  74. audit trail, and raises `GovernanceDenied` when the policy blocks the action.
  75. ```yaml
  76. # policy.yaml
  77. apiVersion: governance.toolkit/v1
  78. name: production-policy
  79. default_action: allow
  80. rules:
  81. - name: block-destructive
  82. condition: "action.type in ['drop', 'delete', 'truncate']"
  83. action: deny
  84. description: "Destructive operations require human approval"
  85. - name: require-approval-for-send
  86. condition: "action.type == 'send_email'"
  87. action: require_approval
  88. approvers: ["security-team"]
  89. ```
  90. ```python
  91. >>> safe_tool(action="read", table="users")
  92. {'table': 'users', 'rows': 42}
  93. >>> safe_tool(action="drop", table="users")
  94. GovernanceDenied: Action denied by policy rule 'block-destructive':
  95. Destructive operations require human approval
  96. ```
  97. Or use the full `AgentControl` API for programmatic control:
  98. <details>
  99. <summary><b>AgentControl example</b></summary>
  100. ```python
  101. from agent_control_specification import AgentControl
  102. runtime = AgentControl.from_path(str("manifest.yaml"))
  103. result = runtime.evaluate(
  104. "input",
  105. {
  106. "envelope": {"agent_id": "example-agent"},
  107. "input": {"body": {"action": "web_search", "params": {}}},
  108. },
  109. )
  110. print(result.verdict)
  111. runtime.close()
  112. ```
  113. [Run the complete ACS email-tool example](examples/acs-email-tool).
  114. </details>
  115. <details>
  116. <summary><b>TypeScript / .NET / Rust / Go examples</b></summary>
  117. **TypeScript**
  118. ```typescript
  119. import { PolicyEngine } from "@microsoft/agent-governance-sdk";
  120. const engine = new PolicyEngine([
  121. { action: "web_search", effect: "allow" },
  122. { action: "shell_exec", effect: "deny" },
  123. ]);
  124. engine.evaluate("web_search"); // "allow"
  125. engine.evaluate("shell_exec"); // "deny"
  126. ```
  127. **.NET**
  128. ```csharp
  129. using AgentGovernance;
  130. using AgentGovernance.Extensions.ModelContextProtocol;
  131. using AgentGovernance.Policy;
  132. var kernel = new GovernanceKernel(new GovernanceOptions
  133. {
  134. PolicyPaths = new() { "policies/default.yaml" },
  135. });
  136. var result = kernel.EvaluateToolCall("did:mesh:agent-1", "web_search",
  137. new() { ["query"] = "latest AI news" });
  138. // MCP server integration
  139. builder.Services.AddMcpServer()
  140. .WithGovernance(options => options.PolicyPaths.Add("policies/mcp.yaml"));
  141. ```
  142. **Rust**
  143. ```rust
  144. use agent_governance::{AgentMeshClient, ClientOptions};
  145. let client = AgentMeshClient::new("my-agent").unwrap();
  146. let result = client.execute_with_governance("data.read", None);
  147. assert!(result.allowed);
  148. ```
  149. **Go**
  150. ```go
  151. import agentmesh "github.com/microsoft/agent-governance-toolkit/agent-governance-golang"
  152. client, _ := agentmesh.NewClient("my-agent",
  153. agentmesh.WithPolicyRules([]agentmesh.PolicyRule{
  154. {Action: "data.read", Effect: agentmesh.Allow},
  155. {Action: "*", Effect: agentmesh.Deny},
  156. }),
  157. )
  158. result := client.ExecuteWithGovernance("data.read", nil)
  159. ```
  160. </details>
  161. CLI tools:
  162. ```bash
  163. agt doctor # check installation
  164. agt verify # OWASP compliance check
  165. agt verify --evidence ./agt-evidence.json --strict # fail CI on weak evidence
  166. agt red-team scan ./prompts/ --min-grade B # prompt injection audit
  167. agt lint-policy policies/ # validate policy files
  168. ```
  169. Full walkthrough: [quickstart.md](docs/quickstart.md) -- zero to governed agents in 5 minutes.
  170. 🌍 Also in: [Español](docs/i18n/quickstart.es.md) | [日本語](docs/i18n/quickstart.ja.md) | [简体中文](docs/i18n/quickstart.zh-CN.md) | [한국어](docs/i18n/quickstart.ko.md)
  171. ---
  172. ## How It Works
  173. ```
  174. Agent ──► Policy Engine ──► Identity ──► Audit Log
  175. (YAML/OPA/Cedar) (SPIFFE/DID/mTLS) (Tamper-evident)
  176. │ │
  177. ├── Allowed ──► Tool executes │
  178. └── Denied ──► GovernanceDenied │
  179. ▼
  180. Decision Record
  181. ```
  182. Every layer is optional. Start with `govern()` and add layers as your risk profile grows. Most teams run policy enforcement + audit logging and never need the full stack.
  183. ---
  184. ## Packages
  185. | Package | Description |
  186. |---------|-------------|
  187. | [**Agent OS**](agent-governance-python/agent-os/) | Policy engine, agent lifecycle, governance gate |
  188. | [**Agent Control Specification**](policy-engine/) ([README](policy-engine/README.md)) | Stateless, deterministic, fail-closed policy decision runtime (Rust core) backing the AGT policy layer |
  189. | [**Agent Mesh**](agent-governance-python/agent-mesh/) | Agent discovery, routing, and trust mesh |
  190. | [**Agent Runtime**](agent-governance-python/agent-runtime/) | Execution sandboxing with four privilege rings |
  191. | [**Agent SRE**](agent-governance-python/agent-sre/) | Kill switch, SLO monitoring, chaos testing |
  192. | [**Agent Compliance**](agent-governance-python/agent-compliance/) | OWASP verification, policy linting, integrity checks |
  193. | [**Agent Marketplace**](agent-governance-python/agent-marketplace/) | Plugin governance and trust scoring |
  194. | [**Agent Lightning**](agent-governance-python/agent-lightning/) | RL training governance with violation penalties |
  195. | [**Agent Hypervisor**](agent-governance-python/agent-hypervisor/) | Execution audit, delta engine, in-memory commitment tracking, command denylist enforcement |
  196. ### Additional Capabilities
  197. | Capability | Description |
  198. |---|---|
  199. | **MCP Security Gateway** | Tool poisoning detection, drift monitoring, typosquatting, hidden instruction scanning ([Spec](docs/specs/MCP-SECURITY-GATEWAY-1.0.md)) |
  200. | **Shadow AI Discovery** | Find unregistered agents across processes, configs, and repos ([Discovery](agent-governance-python/agent-discovery/)) |
  201. | **Governance Dashboard** | Real-time fleet visibility for health, trust, and compliance ([Dashboard](examples/demos/governance-dashboard/)) |
  202. | **PromptDefense Evaluator** | 12-vector prompt injection audit ([Evaluator](agent-governance-python/agent-compliance/src/agent_compliance/prompt_defense.py)) |
  203. | **Contributor Reputation** | PR/issue author screening for social engineering. Reusable GitHub Action ([Action](.github/actions/contributor-check/)) |
  204. ---
  205. ## Install
  206. | Language | Package | Command |
  207. |----------|---------|---------|
  208. | **Python** | [`agent-governance-toolkit`](https://pypi.org/project/agent-governance-toolkit/) | `pip install "agent-governance-toolkit[full]"` |
  209. | **TypeScript** | [`@microsoft/agent-governance-sdk`](agent-governance-typescript/) | `npm install @microsoft/agent-governance-sdk` |
  210. | **Copilot CLI** | [`@microsoft/agent-governance-copilot-cli`](agent-governance-copilot-cli/) | `npx @microsoft/agent-governance-copilot-cli install` |
  211. | **Claude Code** | [`@microsoft/agent-governance-claude-code`](agent-governance-claude-code/) | `claude --plugin-dir ./agent-governance-claude-code` |
  212. | **Codex CLI** | [`@microsoft/agent-governance-codex-cli`](agent-governance-codex-cli/) | `node agent-governance-codex-cli/bin/agt-codex.mjs install` |
  213. | **OpenCode** | [`@microsoft/agent-governance-opencode`](agent-governance-opencode/) | `npm install @microsoft/agent-governance-opencode` |
  214. | **.NET** | [`Microsoft.AgentGovernance`](https://www.nuget.org/packages/Microsoft.AgentGovernance) | `dotnet add package Microsoft.AgentGovernance` |
  215. | **.NET MCP** | `Microsoft.AgentGovernance.Extensions.ModelContextProtocol` | `dotnet add package Microsoft.AgentGovernance.Extensions.ModelContextProtocol` |
  216. | **Rust** | [`agent-governance`](https://crates.io/crates/agent-governance) | `cargo add agent-governance` |
  217. | **Go** | [`agent-governance-toolkit`](agent-governance-golang/) | `go get github.com/microsoft/agent-governance-toolkit/agent-governance-golang` |
  218. All five language SDKs implement core governance (policy, identity, trust, audit). Python has the full stack. Copilot CLI and Claude Code are first-party developer surfaces built on the TypeScript SDK.
  219. See **[Language Package Matrix](docs/PACKAGE-FEATURE-MATRIX.md)** for detailed per-language coverage.
  220. <details>
  221. <summary><b>Python distributions (v4.1.0 — consolidated)</b></summary>
  222. As of v4.1.0, 45 packages have been consolidated into 5 top-level distributions:
  223. | Distribution | PyPI | What's included |
  224. |--------------|------|-----------------|
  225. | `agent-governance-toolkit-core` | [`agent-governance-toolkit-core`](https://pypi.org/project/agent-governance-toolkit-core/) | Policy engine, capability model, audit, MCP gateway, zero-trust identity, trust scoring, A2A/MCP/IATP bridges |
  226. | `agent-governance-toolkit-runtime` | [`agent-governance-toolkit-runtime`](https://pypi.org/project/agent-governance-toolkit-runtime/) | Privilege rings, saga orchestration, termination control, execution plan validation, command denylist enforcement |
  227. | `agent-governance-toolkit-sre` | [`agent-governance-toolkit-sre`](https://pypi.org/project/agent-governance-toolkit-sre/) | SLOs, error budgets, chaos engineering, circuit breakers |
  228. | `agent-governance-toolkit-cli` | [`agent-governance-toolkit-cli`](https://pypi.org/project/agent-governance-toolkit-cli/) | `agt` CLI, OWASP verification, integrity checks, policy linting |
  229. | `agent-governance-toolkit[full]` | [`agent-governance-toolkit`](https://pypi.org/project/agent-governance-toolkit/) | Meta-package installing all of the above |
  230. Previous package names (`agent-os-kernel`, `agentmesh-platform`, `agentmesh-runtime`, `agent-sre`, `agent-discovery`, `agent-hypervisor`, `agentmesh-marketplace`, `agentmesh-lightning`) remain installable as stub packages that redirect to the consolidated distributions.
  231. </details>
  232. ### Prerequisites
  233. - **Python**: 3.10+
  234. - **Node.js**: 18+ / npm 9+ (TypeScript SDK)
  235. - **.NET**: 8+
  236. - **Go**: 1.25+
  237. - **Rust**: 1.70+
  238. - **Optional**: `AZURE_CLIENT_ID`, `AZURE_TENANT_ID`, `AZURE_CLIENT_SECRET` for Azure-integrated features
  239. ---
  240. ## Framework Support
  241. | Framework | Integration |
  242. |-----------|-------------|
  243. | [**Microsoft Agent Framework**](https://github.com/microsoft/agent-framework) | Native Middleware |
  244. | [**Semantic Kernel**](https://github.com/microsoft/semantic-kernel) | Native (.NET + Python) |
  245. | [AutoGen](https://github.com/microsoft/autogen) | Adapter |
  246. | [LangGraph](https://github.com/langchain-ai/langgraph) / [LangChain](https://github.com/langchain-ai/langchain) | Adapter |
  247. | [CrewAI](https://github.com/crewAIInc/crewAI) | Adapter |
  248. | [OpenAI Agents SDK](https://github.com/openai/openai-agents-python) | Middleware |
  249. | Claude Code | Governance plugin package |
  250. | [Google ADK](https://github.com/google/adk-python) | Adapter |
  251. | [LlamaIndex](https://github.com/run-llama/llama_index) | Middleware |
  252. | [Haystack](https://github.com/deepset-ai/haystack) | Pipeline |
  253. | [Mastra](https://github.com/mastra-ai/mastra) | Adapter |
  254. | [Dify](https://github.com/langgenius/dify) | Plugin |
  255. | [Azure AI Foundry](https://learn.microsoft.com/azure/ai-studio/) | Deployment Guide |
  256. | GitHub Copilot CLI | Governance installer |
  257. Full list: [Framework Integrations](agent-governance-python/agentmesh-integrations/) · [Quickstart Examples](examples/quickstart/)
  258. ---
  259. ## Examples
  260. | Example | Framework | What it demonstrates |
  261. |---------|-----------|----------------------|
  262. | [acs-email-tool](examples/acs-email-tool) | Framework-neutral ACS host | Snapshot, verdict, transform, deny, and host enforcement |
  263. | [acs-atr-annotator](examples/acs-atr-annotator) | ACS custom policy | Independent threat-rule annotations with fail-closed decisions |
  264. | [openai-agents-governed](examples/openai-agents-governed) | OpenAI Agents SDK | Policy-gated tool calls with trust tiers |
  265. | [crewai-governed](examples/crewai-governed) | CrewAI | Multi-agent governance with role-based policies |
  266. | [smolagents-governed](examples/smolagents-governed) | HuggingFace smolagents | Lightweight agent governance |
  267. | [maf-integration](examples/maf-integration) | MAF | Microsoft Agent Framework integration |
  268. | [mcp-trust-verified-server](examples/mcp-trust-verified-server) | MCP | Trust-verified MCP server implementation |
  269. | [governance-dashboard](examples/demos/governance-dashboard) | Streamlit | Real-time fleet visibility dashboard |
  270. | [k8s-agent-sandbox-governed](examples/k8s-agent-sandbox-governed) | kubernetes-sigs/agent-sandbox | Pre-dispatch command policy on top of Kubernetes-native execution isolation |
  271. ---
  272. ## Specifications
  273. Every major component has a formal RFC 2119 specification with conformance tests. These specs define the behavioral contract: what implementations MUST, SHOULD, and MAY do.
  274. | Specification | Scope | Tests |
  275. |---|---|---|
  276. | [Agent OS Policy Engine](docs/specs/AGENT-OS-POLICY-ENGINE-1.0.md) | Native runtime integration and fail-closed semantics | -- |
  277. | [Agent Control Specification](policy-engine/spec/SPECIFICATION.md) | Stateless intervention-point policy runtime, verdicts, transform, fail-closed | -- |
  278. | [AgentMesh Identity and Trust](docs/specs/AGENTMESH-IDENTITY-TRUST-1.0.md) | Credentials, trust scoring, delegation chains | 135 |
  279. | [Agent Hypervisor Execution Control](docs/specs/AGENT-HYPERVISOR-EXECUTION-CONTROL-1.0.md) | Privilege rings, saga orchestration, kill switch | 80 |
  280. | [AgentMesh Trust and Coordination](docs/specs/AGENTMESH-TRUST-COORDINATION-1.0.md) | Peer trust negotiation, mesh-wide policy | 62 |
  281. | [Agent SRE Governance](docs/specs/AGENT-SRE-GOVERNANCE-1.0.md) | SLOs, error budgets, chaos, circuit breakers | 111 |
  282. | [MCP Security Gateway](docs/specs/MCP-SECURITY-GATEWAY-1.0.md) | Tool poisoning, drift detection, hidden instructions | 127 |
  283. | [Agent Lightning Fast-Path](docs/specs/AGENT-LIGHTNING-FAST-PATH-1.0.md) | RL training governance, violation penalties | 100 |
  284. | [Framework Adapter Contract](docs/specs/FRAMEWORK-ADAPTER-CONTRACT-1.0.md) | Native framework mediation contract | -- |
  285. | [Audit and Compliance](docs/specs/AUDIT-COMPLIANCE-1.0.md) | Merkle audit, compliance mapping, Decision BOM | 157 |
  286. | [AgentMesh Wire Protocol](docs/specs/AGENTMESH-WIRE-1.0.md) | Message format, routing, serialization | -- |
  287. **992 conformance tests** ensure code stays aligned to specs. [29 Architecture Decision Records](docs/adr/) document why.
  288. ---
  289. ## Standards Compliance
  290. | Standard | Coverage |
  291. |----------|----------|
  292. | [OWASP Agentic AI Top 10](docs/compliance/owasp-agentic-top10-architecture.md) | All ASI risk categories mapped with deterministic controls |
  293. | [NIST AI RMF 1.0](docs/compliance/nist-ai-rmf-alignment.md) | Full GOVERN, MAP, MEASURE, MANAGE alignment |
  294. | [EU AI Act](docs/compliance/) | Compliance mapping with automated evidence |
  295. | [SOC 2](docs/compliance/soc2-mapping.md) | Control mapping with audit trail export |
  296. | [AARM Extended](https://aarm.dev/builders/agent-governance-toolkit-microsoft) | All R1–R9 requirements satisfied; verified Jun 14, 2026 |
  297. | [ATF](https://agentictrustframework.ai/ecosystem) | All five elements mapped: Agent Mesh (identity), Agent OS (policy), Agent Compliance (governance), Agent Runtime (sandboxing), Agent SRE (incident response) |
  298. ---
  299. ## Security
  300. AGT enforces governance at the application middleware layer, not at the OS kernel level. The policy engine and agents share the same process boundary.
  301. **Production recommendation:** Run each agent in a separate container for OS-level isolation. See [Architecture: Security Boundaries](docs/ARCHITECTURE.md).
  302. | Tool | Coverage |
  303. |------|----------|
  304. | CodeQL | Python + TypeScript SAST |
  305. | Gitleaks | Secret scanning on PR/push/weekly |
  306. | ClusterFuzzLite | 7 fuzz targets (policy, injection, MCP, sandbox, trust) |
  307. | Dependabot | 13 ecosystems |
  308. | OpenSSF Scorecard | Weekly scoring + SARIF upload |
  309. See [Known Limitations](docs/LIMITATIONS.md) for honest design boundaries and recommended layered defense.
  310. ---
  311. ## Documentation
  312. | Category | Links |
  313. |----------|-------|
  314. | **Getting Started** | [Quick Start](docs/quickstart.md) · [Tutorials](docs/tutorials/) (60+) · [FAQ](docs/FAQ.md) |
  315. | **Architecture** | [System Design](docs/ARCHITECTURE.md) · [Threat Model](docs/security/threat-model.md) · [ADRs](docs/adr/) (29) |
  316. | **Specifications** | [All Specs](docs/specs/) (10 formal specs, 992 conformance tests) |
  317. | **API Reference** | [Agent OS](agent-governance-python/agent-os/README.md) · [AgentMesh](agent-governance-python/agent-mesh/README.md) · [Agent SRE](agent-governance-python/agent-sre/README.md) |
  318. | **Compliance** | [OWASP](docs/compliance/owasp-agentic-top10-architecture.md) · [EU AI Act](docs/compliance/) · [NIST AI RMF](docs/compliance/nist-ai-rmf-alignment.md) · [SOC 2](docs/compliance/soc2-mapping.md) · [AARM Extended](https://aarm.dev/builders/agent-governance-toolkit-microsoft) · [ATF](https://agentictrustframework.ai/ecosystem) |
  319. | **Deployment** | [Azure](docs/deployment/README.md) · [AWS](docs/deployment/README.md) · [GCP](docs/deployment/README.md) · [Docker Compose](docs/deployment/README.md) |
  320. | **Extensions** | [VS Code](agent-governance-typescript/agent-os-vscode/) · [Framework Integrations](agent-governance-python/agentmesh-integrations/) |
  321. ---
  322. ## Contributing
  323. [Contributing Guide](CONTRIBUTING.md) · [Community](docs/COMMUNITY.md) · [Discord](https://discord.gg/TxMRqY3pFr) · [Security Policy](SECURITY.md) · [Changelog](CHANGELOG.md)
  324. **Using AGT?** Add your organization to [ADOPTERS.md](docs/ADOPTERS.md).
  325. ## Governance
  326. | Document | Purpose |
  327. |----------|---------|
  328. | [GOVERNANCE.md](GOVERNANCE.md) | Decision-making, roles, contributor ladder |
  329. | [CHARTER.md](docs/CHARTER.md) | Technical charter (LF Projects format) |
  330. | [MAINTAINERS.md](MAINTAINERS.md) | Maintainers and organizations |
  331. | [SECURITY.md](SECURITY.md) | Vulnerability reporting and response SLAs |
  332. | [CODE_OF_CONDUCT.md](CODE_OF_CONDUCT.md) | Microsoft Open Source Code of Conduct |
  333. | [ANTITRUST.md](ANTITRUST.md) | Competition law guidelines for participants |
  334. | [TRADEMARKS.md](TRADEMARKS.md) | Trademark usage policy |
  335. ## Important Notes
  336. If you use the Agent Governance Toolkit to build applications that operate with third-party agent frameworks or services, you do so at your own risk. We recommend reviewing all data being shared with third-party services and being cognizant of third-party practices for retention and location of data.
  337. ## Official Sources
  338. The only official sources for the Agent Governance Toolkit are:
  339. | Resource | Location |
  340. |----------|----------|
  341. | **Source code** | [github.com/microsoft/agent-governance-toolkit](https://github.com/microsoft/agent-governance-toolkit) |
  342. | **Documentation** | [microsoft.github.io/agent-governance-toolkit](https://microsoft.github.io/agent-governance-toolkit/) |
  343. | **Python packages** | `https://pypi.org/user/agentgovtoolkit/` |
  344. | **npm packages** | `@microsoft/agent-governance-sdk` on [npmjs.com](https://www.npmjs.com/) |
  345. | **NuGet packages** | `Microsoft.AgentGovernance.*` on [nuget.org](https://www.nuget.org/) |
  346. | **Rust crates** | `agent-governance`, `agent-governance-mcp` on [crates.io](https://crates.io/) |
  347. The project team does not maintain or endorse any third-party websites,
  348. packages, or documentation sites claiming to be official. If you encounter a
  349. suspicious site or package using the Agent Governance Toolkit name, please
  350. report it through the channels described in [SECURITY.md](SECURITY.md).
  351. ## License
  352. This project is licensed under the [MIT License](LICENSE).
  353. ## Trademarks
  354. This project may contain trademarks or logos for projects, products, or services. Authorized use of Microsoft
  355. trademarks or logos is subject to and must follow
  356. [Microsoft's Trademark & Brand Guidelines](https://www.microsoft.com/en-us/legal/intellectualproperty/trademarks/usage/general).
  357. Use of Microsoft trademarks or logos in modified versions of this project must not cause confusion or imply Microsoft sponsorship.
  358. Any use of third-party trademarks or logos are subject to those third-party's policies.

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.