The eval-runner workflow from the .github/workflows directory is too big. Read it all carefully, and note how it's doing two different things: integration test evaluation and SWE-Bench evaluation. Let's split it:
1) find the section named "Run integration test evaluation" and the necessary sections before and after
2) create a new workflow named integration-runner
3) define it as to work on a PR with the label 'integration-test' trigger
4) move the integration test evaluation section to the new file and copy the prerequisites.
IMPORTANT:
the action should run only when labeled or manually triggered.
The eval-runner workflow from the .github/workflows directory is too big. Read it all carefully, and note how it's doing two different things: integration test evaluation and SWE-Bench evaluation. Let's split it: 1) find the section named "Run integration test evaluation" and the necessary sections before and after 2) create a new workflow named integration-runner 3) define it as to work on a PR with the label 'integration-test' trigger 4) move the integration test evaluation section to the new file and copy the prerequisites.
IMPORTANT: