Dashboard

GitHub agentic autofix is GA: what it fixes and costs

GitHub's agentic autofix for code scanning alerts is generally available, with AI Scan for pull requests. Who gets it, how it works and how it is billed.

Cecilia Iona
Cecilia Iona
Senior Editor, AI & Product
17 October 20261 min read

GitHub made agentic autofix for code scanning alerts generally available on 15 October 2026, and on the same day made AI Scan for pull requests generally available. Autofix lets Copilot work on a security alert across your codebase and open a draft pull request with the fix. AI Scan checks pull requests in languages CodeQL does not yet cover. Both are paid features aimed at organisations on GitHub's security products, so most solo builders will not get them by default.

Details below come from GitHub's changelog entries for agentic autofix and AI Scan.

What agentic autofix does

Once you assign a code scanning alert to Copilot, it explores the relevant files, writes a fix, reruns CodeQL to check the fix, and iterates if the check fails. It then opens a draft pull request containing a fix summary, the reason the change closes the alert, and the validation steps it ran. You can ask for changes by commenting on the pull request or from the Agents tab.

Practical details from the announcement:

  • It replaces the free Generate Fix option with an Assign to Copilot button.

  • It works on any alert on the default branch. Alerts found only on pull requests or other branches cannot be assigned.

  • You can select up to 25 alerts in the security alerts list or a security campaign and fix them together in one pull request.

  • There is an API route: set assignees to ["copilot-swe-agent[bot]"] in the Update a Code Scanning Alert endpoint.

That bulk option deserves care. One pull request that closes 25 alerts is quick to create and slow to review properly. How to review a large AI-generated pull request efficiently applies directly, and so does checking the diff the way this guide to reviewing an agent's git diff describes.

What AI Scan adds

AI Scan complements CodeQL by reviewing pull request changes for languages and frameworks CodeQL does not yet cover, including PHP, Shell and Bash, Terraform (HCL), Dockerfiles, JSP for Java and Blazor for C#. It runs automatically on pull requests when code scanning and the repository's AI Scan setting are both on.

Findings carry an "AI" marker so they are distinguishable from CodeQL alerts, and can include an explanation and a suggested fix. Two limits matter:

  • Findings are advisory. They do not fail the code scanning check and cannot block a merge through ruleset requirements.

  • It does not run on pull requests from forks or on pull requests created by Dependabot.

A Copilot licence is not required for AI Scan.

Who gets it, and what it costs

This is the part that decides whether you care.

  • Autofix is available to GitHub Code Security and GitHub Advanced Security customers on GitHub Enterprise Cloud, including Enterprise Cloud with data residency. Three things have to line up: an enterprise administrator enables the "Agentic autofix" policy, an organisation or repository administrator turns it on in Code Security settings, and a Copilot billing method is defined.

  • AI Scan is available on public repositories owned by organisations or personal accounts. Private and internal organisation repositories need GitHub Code Security or Advanced Security.

  • Billing for both uses Code Security AI Credits. Autofix also uses GitHub Actions minutes. You can set SKU-level budgets for the credits.

GitHub's pages do not state a price per credit, the credit cost of a single fix, or the programming languages autofix supports. Do not budget from this announcement alone.

What to do with this

  • Solo or small team on a public repository: AI Scan is the one you can plausibly use today. Treat its findings as a second opinion, not a gate.

  • Organisation on Enterprise Cloud with Advanced Security: ask the enterprise admin whether the autofix policy is on, and set a credit budget before the first bulk run.

  • Everyone else: nothing to switch on. If you want a sense of what AI security review can and cannot find, can AI find security vulnerabilities in your code is a grounded starting point, and the wider landscape of assistants is in the AI coding tools guide.

One more habit worth keeping: an autofix that turns the alert green has not proven the code is safe. It has proven the scanner stopped complaining. How to catch an AI coding agent introducing a vulnerability covers the failure mode.

How did this land?

About the author

Cecilia Iona
Cecilia Iona

Senior Editor, AI & Product

Cecilia leads the Swarmz editorial desk. She has spent a decade turning complex AI and product topics into writing people actually finish, and she owns the blog's quality bar.

Share

Get the next post in your inbox

One email a month. Product updates, engineering posts, and the best of Built with Swarmz.

I agree to receive emails about AI building tips and Swarmz product news. Unsubscribe any time.