· 6 min read

GitHub Will Pause Dependabot Alerts on Repos With No Push in a Year Starting December 8. Your Live Side Projects Are the Ones That Go Quiet.

GitHub posted a changelog entry on November 10 saying that, starting December 8, it will turn off Dependabot alerts on most repositories with no activity in the past year. Existing alerts get archived, not deleted. For GitHub, this is sensible housekeeping across millions of abandoned repos. For a solo operator, it hits exactly the wrong set of repos: the finished ones that still run in production and that you haven't touched since last summer.

What GitHub is changing

The rule is simple. If a repository has had no activity in the past year, Dependabot alerts get switched off on December 8, and any open alerts move to an archive you can restore later.

The definition of activity is narrow. GitHub lists three things: pushing to any branch or tag, creating the repository, and turning on Dependabot alerts. Opening issues, merging someone else's pull request through the UI without a push you made, or a deploy pipeline pulling the code aren't in that list. The changelog doesn't explicitly rule them out, but it presents those three as the set, so I'd assume nothing else counts.

Coming back is easy. Any push re-enables alerts automatically, or you can turn them on again in the repository's security settings. Either way the archived alerts come back and the one-year clock restarts.

There are two exemptions. Repositories covered by an enforced security configuration that turns Dependabot alerts on aren't paused, and neither are repos in enterprises whose policy stops repository admins from changing the alert setting. The change applies to all plans on github.com and to GitHub Enterprise Cloud with data residency. It doesn't apply to GitHub Enterprise Server.

The changelog only talks about alerts. It doesn't say what happens to Dependabot security update pull requests or version update pull requests on a paused repo. Since security update PRs are generated from alerts, I'd expect those to stop too, but GitHub hasn't said so.

Why solo operators get hit harder than teams

On a team, the repos that run production get pushed to every week. In a one-person business, the pattern is different. You ship a landing page, a small SaaS, a client site, a Chrome extension backend, and then you leave it alone because it works. My own GitHub has a dozen repos like that. Some are dead. Several are deployed, take payments, or hold customer email addresses.

Those are the repos where a Dependabot alert email is the only thing telling you a dependency just got a critical CVE. When the alerts stop, nothing else in your setup replaces them unless you built it. And "no activity for a year" is also a good description of a repo with a year's worth of unpatched dependencies.

The 15-minute inventory

Here's what I'd do before December 8.

First, list repos by last push, oldest first. With the GitHub CLI:

gh repo list --limit 200 --json name,pushedAt,isArchived \
  --jq '.[] | select(.isArchived == false) | "\(.pushedAt)  \(.name)"' | sort

Anything with a pushedAt older than about mid-December 2025 is a candidate for the pause, because it will have had no push in the year before December 8.

Second, mark each candidate as one of three things: deployed and important, deployed but disposable, or dead. Be honest about the middle category. A marketing site with a contact form is deployed and important if the form writes to a database.

Third, act on each group:

  1. Dead repos: archive them on GitHub. That's cleaner than letting alerts lapse silently, and it makes the next inventory shorter.
  2. Deployed and important: either push something real (update the lockfile, bump a dependency, fix the README) or re-enable alerts in settings on or after December 8. A real dependency update is better than an empty commit, because it also fixes whatever's been piling up.
  3. Deployed but disposable: decide whether you'd act on an alert anyway. If not, take it offline instead.

If you have a free or paid organization for your projects, look at security configurations. An enforced configuration that turns on Dependabot alerts exempts its repos from the pause, and it's one setting instead of a dozen manual re-enables.

Where I could be wrong

There's a fair argument that this is the right default and I'm over-reacting. If you haven't pushed to a repo in a year, you probably haven't been acting on its alerts either, and a backlog of 40 ignored alerts isn't protection. GitHub is also clear that nothing is deleted and that one push brings everything back.

I agree with that for most repos. My worry is the specific case where a repo is quiet because the product is finished, and the alert email is the only monitoring it has. For those, quiet is a bug. The fix takes minutes, so I'd still do the inventory.

What I'd actually do

This weekend: run the gh repo list command, sort the output, and archive the dead repos. Then, for each deployed repo older than a year, open a branch, run your package manager's update command, test it, and push. That clears the backlog and resets the clock in one move.

Longer term, I'd put a recurring reminder on the calendar for every December: same command, same three groups. A yearly dependency refresh on every live repo is cheap insurance, and now it also keeps GitHub's alerts switched on.

Author

Sources

Newsletter

The week, in one email.

The weekly summary of what happened and what it means for people who build alone. No spam, unsubscribe any time.

Form not showing? Subscribe on this page instead.

Related Posts