AI news · 2 min read
GitHub expands AI security checks, but a finding will not block a merge
GitHub’s October 15 releases add advisory AI Scan findings and agentic repair of default-branch alerts. Access, review and AI-credit billing differ.
Announcement date:
Affiliate disclosure: Some links may earn CreatorStack a commission at no extra cost to you.
At a glance
GitHub made AI Scan and agentic autofix generally available on October 15. AI Scan gives advisory pull-request findings that cannot block merges. Agentic autofix handles eligible default-branch alerts and proposes draft pull requests. Their access requirements differ, and both consume Code Security AI Credits.
An AI warning in a pull request does not necessarily stop the change from shipping. GitHub’s October 15, 2026 releases make AI Scan and agentic autofix generally available, but they serve different stages of security work: spotting a possible problem in a proposed change, and preparing a repair for an existing alert.
First, understand what the finding can enforce
AI Scan complements CodeQL in areas such as PHP, Shell, Terraform configuration and Dockerfiles. With the required repository settings enabled, its findings appear beside CodeQL results with an AI label. They are advisory: they do not fail the scanning check or become a merge-blocking ruleset requirement. Fork and Dependabot pull requests are excluded. AI Scan announcement.
For a creator maintaining a small website, that means an extra review signal still needs someone to read it. A suggested fix should explain the risk, preserve the intended behaviour and pass the project’s checks. A green merge button alone does not answer those questions.
Repair works from an eligible default-branch alert
Agentic autofix explores the repository, proposes changes, reruns CodeQL and opens a draft pull request for review. It supports default-branch alerts; an alert found only on another branch cannot be assigned through this workflow. Up to 25 alerts can be grouped for one proposed repair. Autofix release.
Access is narrower than AI Scan: the agentic repair release requires an eligible Code Security or Advanced Security customer on Enterprise Cloud, administrator enablement and a Copilot billing method. The result is a proposal to examine, not a reason to skip your own acceptance checks.
Public availability still has a usage bill
AI Scan is available for public repositories; private and internal organization repositories require the relevant security product. A Copilot licence is not required for AI Scan. Both features use Code Security AI Credits, including on public repositories, and agentic autofix also uses Actions minutes. Review the billed account’s allowance and budget before enabling them broadly. Security billing rules.
Start with one understood alert and inspect the proposed diff and validation. Our AI output evaluation guide and coverage of local Copilot sandboxing explain two useful boundaries around that review.
Common questions
Questions people ask.
- Is AI Scan free just because my repository is public?
- Public repository availability does not remove AI-credit billing. GitHub says AI Scan and agentic autofix consume Code Security AI Credits even where active-committer licences are not required.
Sources & freshness
This article was refreshed on 16 Oct 2026 using the sources below. A source check verifies published information, not output quality or checkout availability in every region.
- github.blog/changelog/2026-10-15-ai-scan-for-pull-requests-generally-available/
- github.blog/changelog/2026-10-15-agentic-autofix-for-code-scanning-alerts-general-availability/
- docs.github.com/en/billing/concepts/product-billing/github-advanced-security
- docs.github.com/en/enterprise-cloud@latest/code-security/concepts/code-scanning/autofix-for-code-scanning
What changed: Original coverage of the October 15 announcement with product, access and practical limits checked October 16.
Found a correction? Let us know.