Skip to content
DevOps

Govern AI agents on GitHub

Apply governance to AI agents that create PRs, manage issues, and push code. The GitHub pack enforces pre-dispatch policy checks on repository operations, requires approvals for sensitive actions, and logs every agent interaction with your codebase.

What this pack does

  • Pre-dispatch policy checks on PR creation and code pushes
  • Branch protection enforcement for agent commits
  • Approval gates for repository-level operations
  • Full audit trail of agent-initiated changes

Use cases

Require approval before an agent merges to main

Block agents from pushing to protected branches

Audit all AI-generated pull requests

Quick setup

  1. 1Install the GitHub pack: cordumctl pack install github
  2. 2Configure a GitHub App or PAT with required scopes
  3. 3Define policy rules for repository operations
  4. 4Enable the pack and test with a dry-run agent

Frequently asked questions

How does Cordum govern GitHub actions?

Cordum evaluates every GitHub action against your policy before execution. The Safety Kernel returns Allow, Deny, or Require Approval decisions, ensuring agents operate within approved boundaries.

Do I need to modify my existing GitHub setup?

No. The Cordum GitHub pack installs as an overlay. It intercepts agent actions at the governance layer without changing your existing GitHub configuration.

What happens if an agent action is denied?

The action is blocked before execution, logged in the audit trail, and optionally triggers an alert. The agent receives a structured denial with the policy reason, so it can adjust its approach.

Ready to govern GitHub?

Other integrations