By Hamza Faisal

GitHub has given open-source maintainers one more month to fix their automations. The company announced on November 2 that it is pushing enforcement of its default block on pull_request_target workflows in public repositories from November 2 to December 2, 2026, after the rollout schedule slipped.

The delay is not a reprieve from the change itself. GitHub will still block the trigger by default on December 2 for public repositories that have no applicable Actions event policy, and it will test the enforcement twice beforehand with scheduled blackouts on November 9 and November 16, from 16:00 to 18:00 UTC each day. During those windows, blocked workflow runs will fail with an error annotation instead of executing.

Why GitHub is blocking this trigger

The trigger at the centre of the change, pull_request_target, runs workflows in the context of a repository's base branch rather than the pull request's code. That gives it access to repository secrets and a privileged token, which is exactly what makes it useful for tasks like labelling pull requests, posting preview deployments, and running checks on contributions from forks.

It is also what makes it dangerous. A workflow that checks out and runs code from an untrusted pull request while holding those privileges can be abused in what the security community calls a pwn request, where malicious code exfiltrates secrets or tampers with the repository. GitHub has been tightening this area all year: in June, version 7 of its checkout action began blocking fork code checkouts by default in untrusted triggers, and in September its workflow execution protections reached general availability with the default blocking rule in evaluate mode.

What changes on December 2

From December 2, public repositories on GitHub Free, Team, and Enterprise Cloud that were using the default pull_request_target policy before general availability will have the trigger blocked unless an event policy explicitly allows it. Private and internal repositories are not affected, nor are GitHub Enterprise Server instances, and any repository that already has an applicable event policy keeps its existing behaviour.

Until enforcement begins, the rule stays in evaluate mode: workflows keep running, but GitHub's policy Insights show which runs would have been blocked. That preview is the point of the delay — maintainers can see exactly what will break before it breaks, rather than discovering it when a contributor's pull request silently fails.

What maintainers should do before November 9

GitHub's advice is to audit before the first blackout. Maintainers should check Insights for workflows that would be blocked, switch to the safer pull_request event wherever the privileged context is not actually needed, and, where pull_request_target is genuinely required, create an event policy that allows it — ideally narrowed to specific workflow files rather than the whole repository.

The stakes are practical rather than theoretical. Labelers, preview deploys, and fork CI pipelines are the workflows most likely to depend on this trigger, and they tend to break quietly: a failed run looks like a flaky check, not a security policy taking effect. The November 9 and 16 blackouts are designed to make that failure loud while there is still time to fix it, so that December 2 arrives as a formality instead of a fire drill.