← Back to all projects

Agent Governance Toolkit

Wrap a tool in a written policy so that every call is checked, logged, and refused when the rules say no

OfficialMIT
Stars
6.5k
Forks
1.1k
Open issues
120
Last commit
30 Oct 2026

What is Agent Governance Toolkit?

The permissions an agent inherits are the ones its credentials carry, which is almost never the set of things it should be allowed to do: a key that can read a table can usually drop it. This toolkit closes that gap at the point of the call. You write the rules as a YAML file — deny anything destructive, require a named group's approval before an email goes out — wrap the tool function, and every call is evaluated, written to an audit trail, and refused with an explanation when a rule blocks it. Alongside that sit an identity scheme for attributing an action to a specific agent and a sandboxing layer for tools you would rather not run in your own process. It is a public preview, and the package layout has already moved once.

What can you do with Agent Governance Toolkit?

  • Govern an existing tool in two lines — Wrap the function you already have and hand it a policy file — the call site does not change, and neither does the agent framework around it.
  • Write the rules as a file, not as code — Policies are YAML with conditions and outcomes, so what an agent may do can be reviewed, versioned and changed without a deployment of the agent.
  • Require a person for the dangerous cases — A rule can demand approval from a named group before an action proceeds, which is the difference between an audit trail and an actual control.
  • Know which agent did what — Actions are attributed to a specific agent identity rather than to a shared service account, so an audit trail can answer the question afterwards.
  • Refuse with a reason — A blocked call raises an error naming the rule that stopped it, so the failure is debuggable instead of arriving as a silent empty result.
  • Reach it from more than Python — The same governance is published for TypeScript and .NET, and as a plugin for Claude Code, so a mixed estate is covered by one set of policies.

Before you choose Agent Governance Toolkit

  • It is published as a public preview and says breaking changes may come before general availability, so pin the version and read the changelog before upgrading.
  • The packaging has already been reorganised once: the base install carries only the compliance command, the governance modules need an extra, and one earlier import path is deprecated with a one-way migration.

Star history

27 Aug to 30 Oct · +366

6.1k6.5k

Frequently asked questions

Is Agent Governance Toolkit free for commercial use?

Agent Governance Toolkit is released under the MIT licence — OSI-approved open source, which permits commercial use.

How can Agent Governance Toolkit be deployed?

Agent Governance Toolkit is available as Self-hosted / Runs locally.

Agent Governance Toolkit
Ask AI
GitHub