Building an AI bug-fixing agent that runs inside your GitHub Actions runner

Description and use

An AI bug‑fixing agent runs inside a self‑hosted GitHub Actions runner, keeping code private and limiting the agent to reading files and creating pull requests, but it cannot run tests.

Problem

Production errors need automated fixes, but existing SaaS tools expose code to external services.

Outcome

The system achieves automated bug fixes while keeping code private and limiting agent blast radius, but cannot run tests and requires PR permissions to be enabled.

Workflow

• Receive error from tracker
• Control plane starts workflow via workflow_dispatch
• Runner checks out repo and runs agent
• Agent explores code, writes fix to new branch
• Agent opens PR
• Runner reports status, PR link, summary, token counts back to control plane

Evidence

The service that coordinates the agent should never be able to read the code.

Source

The agent has four tools: get_file_contents, list_directory, search_code, create_pull_request.

Source

The agent cannot run your tests. It can read a test file and mention it, but it cannot execute anything.

Source