Playbooks · Coding tasks
Secret scanning now catches Actions log leaks
Written November 10, 2026
GitHub's secret scanning now checks Actions workflow logs for leaked secrets. It runs when a workflow finishes, and the alert stays in your list even after the logs are deleted.
The problem
Workflow logs are a quiet place for a secret to end up. A step prints an environment variable, a debug flag echoes a token, a command fails and dumps its arguments — and the secret sits in the log, readable by anyone who can see the run. Code scanning never looked there. Until now, nothing watched the logs at all.
What changed
GitHub's changelog entry, published November 9, says secret scanning now detects secrets in GitHub Actions workflow logs, in addition to issues, wikis, discussions, and pull requests. When a workflow run completes, GitHub scans its logs for provider patterns. Findings appear in the repository's secret scanning alert list, alongside alerts from code and other content, with a link back to the Actions workflow log. Validity checks run for supported secret types, and alerts remain available even after the logs are deleted.
The feature is in public preview. It is available to enterprises on GitHub Enterprise Cloud with GitHub Secret Protection or GitHub Advanced Security, including those with data residency on GHE.com.
How to turn it on
Enterprise owners enable Secret scanning for Actions workflow logs in enterprise settings, under Policies, then Advanced Security, selecting Enabled in the All repositories dropdown. The policy is disabled by default, and it applies only to repositories with secret scanning enabled under those licenses.
Two things to know before you rely on it. Only workflow runs that complete after you enable the feature are scanned; existing logs are not backfilled, so old leaks in old logs will not surface. And alerts from Actions workflow logs do not generate notifications during the preview — check the repository's secret scanning alert list directly instead of waiting for an email.
What it does not catch
Generic patterns, custom patterns, AI-detected secrets, and secrets already masked by Actions are not detected. That leaves real gaps: a password printed as plain text in a log line, or a secret that Actions already masked with asterisks, will not produce an alert. The feature catches provider-pattern secrets — the recognizable shapes of known services — not every secret-shaped string.
What to do when an alert fires
- Rotate the credential first. A secret that appeared in a log may already have been copied. Revoking it closes the hole; deleting the log does not.
- Find how it got there. Follow the alert's link back to the workflow log, find the step that printed it, and fix the step — a debug flag left on, a secret passed on a command line instead of through the environment.
- Check older runs by hand. Runs from before the feature was enabled were never scanned. If a workflow has ever printed secrets, its history is worth a look.
Limits
- This is a public preview for Enterprise Cloud customers with Secret Protection or Advanced Security. It is not available to everyone, and preview behaviour can change.
- Detection covers provider patterns. Generic and custom patterns, AI-detected secrets, and secrets Actions already masked are not detected.
- Only runs that complete after enabling are scanned. Nothing already in your log history is checked.
- An alert means a secret-shaped value was found in a log. It does not mean the secret was used by anyone.
Related playbooks
- Secret scanning now catches Cloudflare Access tokens
- Lock down agent-built GitHub Actions workflows
- GitHub Agentic Workflows are now GA
- Claude agents now dream: memory that consolidates
Sources
- Secret scanning detects secrets in GitHub Actions workflow logs — GitHub Changelog
- Secret scanning — GitHub Enterprise Cloud Docs
Back to: Coding playbooks hub · Also: Apiiro: 1,122 Claude settings files leak secrets · Copilot CLI 1.0.108: index, PDFs, Ctrl+C