Troubleshooting
A job fails after moving
The usual reasons a job that passed on GitHub's runners fails or doesn't start on Skipbox, and the fix for each.
A command isn't found
Skipbox's image has fewer tools than GitHub's ubuntu-latest. Add the tool with its setup action, such as actions/setup-java or actions/setup-go, or install it with sudo apt-get install -y. See Runner image.
It doesn't start
Your account's page shows a Didn't start line with the reason and what to do. Most say: cancel the run on GitHub, then re-run it. When it says the app isn't installed, or Skipbox can't see the repository, install the app again or add the repository on the app's Configure page first. A job that didn't start used none of your trial's job-minutes.
It never leaves the queue
See A job stays queued.
A network step times out
- Mail: TCP port 25 is blocked. Use your mail provider's HTTP API.
- An inbound webhook or tunnel: nothing can connect in to a job's VM.
- A service that allows only listed IP addresses: Skipbox has no fixed address to list.
- Slow calls to a server far from Europe: jobs run in Helsinki.
See Network and egress.
Docker pulls fail with a rate limit
Log in to Docker Hub in the job with docker/login-action before pulling. See Network and egress.
It stops after 6 hours
That's Skipbox's limit per job, the same as on GitHub-hosted runners. See Job time limit.
It needs Arm, Windows or macOS
Skipbox runs Ubuntu on x64 only. Keep that job on GitHub's runners.
Still failing
Write to hello@skipbox.dev with the run's link.
Updated November 9, 2026