Stabilize tests workflow

1. Run the tests to identify unstable or flaky tests.
   • If specific tests are specified, run only those tests; otherwise, run the entire test suite.
2. Investigate the root cause of each unstable test before changing code.
3. Implement fixes to stabilize the tests.
   • Prefer actual API calls over mocks.
4. Re-run the tests to confirm they pass reliably. If they fail, repeat steps 2-4 until they are stable.
5. Commit and push the changes, then run either the open-pr or update-pr skill.