FR
live
tag

#ci-cd

GitLab 19.3.2 closes an unauthenticated arbitrary file read and seventeen more flaws

GitLab shipped versions 19.3.2, 19.2.6 and 19.1.8 on September 10, 2026 to fix eighteen flaws, including an unauthenticated arbitrary file read through the commits API and an insecure deserialization scored CVSS 9.9. Every exposed self-managed instance must be updated without delay, and protected CI/CD variable secrets need a review.

GitHub Actions adds a vulnerability-alerts token and reusable workflow identity

On September 3, 2026, GitHub shipped three GitHub Actions updates: a vulnerability-alerts permission for GITHUB_TOKEN, the job context for reusable workflows, and a runner deprecation API. Swap your broad scopes for the vulnerability-alerts permission and adopt job.workflow_ref in your reusable workflows.

GITHUB_TOKEN gains a dedicated read permission for Dependabot alerts

In early August 2026, GitHub shipped a vulnerability-alerts: read permission that lets the CI token query Dependabot alerts without an over-privileged PAT. Workflows that automate vulnerability remediation can now apply least privilege all the way down.

GitLab patches CVE-2026-18252, command execution via the Duo Claude agent in CI

On August 26, 2026, GitLab shipped a fix for CVE-2026-18252, a flaw in the Duo Claude AI agent that lets an authenticated developer execute arbitrary commands in a CI context through user-controlled configuration. The lesson goes beyond GitLab: an AI agent wired into CI is a new execution surface, and the configuration it consumes is now part of the security boundary.

A GitHub issue with zero repo privileges can run code on Anthropic and Google CI runners — Black Hat 2026 tears apart coding agent trust

On August 5, 2026, Novee Security demonstrated at Black Hat USA that a GitHub issue opened by an account with no write access was enough to execute arbitrary code on the CI runners behind Claude Code, Gemini CLI, and OpenAI Codex repositories. If your CI/CD pipeline executes code from GitHub issues without sandboxing, treat this as a CVE with no patch — yet.

Type at least two characters.

navigate open esc dismiss