GitHub Actions / GitHub Actions

GitHub Actions permission denied publickey

Fix GitHub Actions SSH permission denied errors when a workflow clones, deploys, or pushes with a key.

Category
GitHub Actions
Error signature
Permission denied (publickey)
Quick fix
Store the private key in a secret, add the matching public key to the target repo, and configure known_hosts before SSH commands.
Updated

What this error means

Permission denied (publickey) means the build or deployment failed in a clean automation environment. The cause is usually runtime version, lockfile state, secrets, project root, or deploy permissions.

Common causes

Copy-paste commands

Check local Node version

node --version
npm --version

Reproduce a clean install

rm -rf node_modules
npm ci

Run the production build locally

npm run build

Check GitHub SSH from a runner-like shell

ssh -T git@github.com

Quick fixes

  1. Open the failed log and find the first error line above the stack trace.
  2. Store the private key in a secret, add the matching public key to the target repo, and configure known_hosts before SSH commands.
  3. Check Node version, working directory, lockfile state, and required secrets.
  4. Rerun the job only after committing the config or lockfile change.

Step-by-step troubleshooting

  1. Find the first log line containing Permission denied (publickey).
  2. Check the job Node version and package manager command.
  3. Verify secrets are available for the event type; forked PRs often have restricted secrets.
  4. Compare the workflow working directory with the folder containing package.json.
  5. Run the same install and build commands locally from a clean checkout.

Platform-specific fixes

GitHub Actions

Vercel

Real-world fixes

How to prevent it

FAQ

What should I check first?

Start with the exact Permission denied (publickey) line and the command, request, or workflow step that produced it. In CI/CD, the first useful clue is usually near the first failure line, not the final stack trace.

Can I ignore this error?

No. Treat it as a failed CI/CD step. A temporary bypass may help diagnosis, but the underlying cause should be fixed before shipping or publishing changes.

Why does this work locally but fail elsewhere?

Local machines often have cached credentials, old dependencies, different runtime versions, or network settings that CI and production do not share. Reproduce from a clean shell or clean install when possible.

How do I know the fix worked?

Rerun the smallest command, request, or deployment step that produced Permission denied (publickey). The fix is working when that step completes without the same signature and produces the expected output.