agent-governance-toolkit-cli

What it is and what it does

agent-governance-toolkit-cli is a CLI and observability layer for the Agent Governance Toolkit, a Microsoft project focused on managing and monitoring AI agents. It consolidates three previously separate tools (agent-sre, agt-sandbox, and agentmesh-mcp-trust) into one package, providing command-line interfaces for agent governance, sandbox isolation to constrain agent behavior, and SRE observability via OpenTelemetry integration.

The package depends on agent-governance-toolkit-core for its core logic, croniter for scheduling, pydantic for data validation, PyYAML for configuration, and OpenTelemetry libraries for metrics and tracing. It targets Python 3.11+ and is in Beta development status. Optional extras (docker, mcp) extend functionality for containerized and Model Context Protocol scenarios.

Use it for:

• Set up CLI commands to monitor and manage AI agents in production environments with SRE observability.
• Isolate untrusted or experimental agents in a sandbox to prevent unauthorized system access or resource consumption.
• Integrate agent governance with OpenTelemetry for centralized tracing and metrics collection.
• Migrate from separate agent-sre, agt-sandbox, or agentmesh-mcp-trust installations to a unified distribution.
• Configure agent behavior policies and trust rules via YAML configuration files.
• Schedule agent tasks or health checks using cron-like expressions.

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

Provides CLI tools, SRE observability, and sandbox isolation for managing and governing AI agents, consolidating agent-sre, agt-sandbox, and agentmesh-mcp-trust into a single distribution.

Yes, if you are building or operating AI agents and need governance, sandbox isolation, and observability. The package is actively maintained, has low install friction, and is backed by Microsoft's governance toolkit. It is in Beta, so expect API changes; review the migration guide if upgrading from the older separate packages. No known security vulnerabilities as of 2026-08-14.

Install

agent-governance-toolkit-cli on PyPI

pip install agent-governance-toolkit-cli

uv add agent-governance-toolkit-cli

poetry add agent-governance-toolkit-cli

Installing agent-governance-toolkit-cli

Before you install

Low friction install with a pure-Python wheel. Active maintenance—released 11 days ago with 5929 GitHub stars and a commit history through 2026-08-14. Supports Python 3.11, 3.12, and 3.13.

License in practice

MIT license (permissive). No restrictions on commercial or private use; you may modify and distribute freely under the same license terms.

Quickstart

pip install agent-governance-toolkit-cli
pip install agent-governance-toolkit-cli[docker,mcp]

from agent_governance_toolkit_cli import ...

Requires Python 3.11 or later. Optional extras (docker, mcp) available for extended functionality.

Verify before relying

• What specific CLI commands are available and what each one does
• How sandbox isolation is configured and what threat model it addresses
• Integration details with opentelemetry-api and opentelemetry-sdk for observability
• Whether the package is suitable for production use given its Beta development status

Package facts

Evidence: agent_governance_toolkit_cli-5.0.0-py3-none-any.whl

Keywords: ai-agents, cli, governance, mcp, observability, reliability, sandbox, sre, trust

Tags

More Libraries packages

Further reading