fix-pr-tests

To fix failing GitHub pull request tests, you can use this Skill to fetch failing check logs via GitHub CLI, reproduce the exact failing test locally, and classify the failure as a real bug, stale test, or flake. It then applies the minimal correct fix and re-verifies before pushing.

CI test failures that differ from local runs often stem from environment deltas. This Skill identifies these runner environment differences by reproducing the exact failing test locally, helping you determine whether the issue is a true code bug or a flaky test caused by environment variations.

The best way to diagnose flaky tests in a GitHub PR is to fetch the exact assertion message and source location from the CI run logs, reproduce the test locally to check for environment deltas, and classify the root cause to apply the minimal fix only to the correct side.

Yes, you need GitHub CLI installed and authenticated to use this Skill. It requires GitHub CLI commands like `gh pr checks` and `gh run view` to fetch failing check details, pull exact assertion messages, and identify runner environment deltas for local reproduction.

Yes, you can safely automate fixing stale tests by classifying each failure as a real bug, stale test, or flake first. This Skill applies the minimal change to only the correct side—production code or test file—and re-verifies the relevant tests before pushing to prevent regressions.

When multiple GitHub Actions jobs are red on your pull request, this Skill finds the true failing assertion by pulling job details, reproduces the specific failing test locally, and applies the smallest correct fix to make CI green again without guessing the root cause.