Skip to main content

Critical CVE Patching

Featured

When a critical advisory is published, finds every repository, image and running workload that carries the package, gets the scope approved, opens a patch PR per repository, and reports hourly until nothing vulnerable runs.

How it works​

  1. Runs when: a critical security advisory is published.
  2. Maps exposure from three sources: manifests, container SBOMs and scan findings, and what actually runs in the cluster, including sidecars and CronJobs.
  3. Orders the work: internet-facing workloads first, then tier 1, then tier 2. Orphaned workloads get an owner.
  4. Gets one approval: the scope and order go to your security approver before any pull request opens.
  5. You get: one minimal patch pull request per repository and an hourly status update. The advisory closes when nothing running is on a vulnerable version, not when the pull requests merge.

What it needs​

  • GitHub: read lockfiles, open pull requests
  • Snyk: read container SBOMs
  • Kubernetes: read running workloads
  • Jira and Slack: track and report status