CyberBoar / agent policy control

Your agents.
Your rules.

Define what coding agents can do.
See what they try.

Define a YAML policy for file access, shell commands, and MCP tools. CyberBoar checks each supported tool call through your coding agent's hook interface.

Get started Explore the docs

01FilesKeep sensitive paths off limits.

02CommandsCheck what the agent tries to run.

03MCP toolsDecide which tools may be called.

Start local. Deploy with confidence.

Install CyberBoar, configure your agent, and manage policies across your team.

01 / Setup

Install and test

Install for one Claude Code project, block a fake secret, and learn how to roll back.

Get started 02 / Policy

Define your rules

Define file, command, and MCP rules, then validate and explain policy decisions.

Read the policy reference 03 / Operations

Deploy across a team

Distribute signed policies, collect audit records, and choose your deployment model.

Read the operations guide

Small policy. Clear boundary.

Make your first test
a harmless one.

The getting-started guide uses a fake secret and one Claude Code project. Validate the policy, install the local hook, then check that an actual agent call is blocked.

Follow the walkthrough

apiVersion: cyberboar.dev/v1
kind: Policy
metadata:
  name: starter
spec:
  defaults:
    posture: deny-listed
  files:
    deny:
      - "**/.env*"

DENYread .envMatched file policy