en un clic
babysit-pr
// Get a pull request to green CI. Diagnose and fix CI failures, push fixes, re-trigger CI via the "Run CICD" label, and repeat until all checks pass. Does not post comments — this is a local developer tool.
// Get a pull request to green CI. Diagnose and fix CI failures, push fixes, re-trigger CI via the "Run CICD" label, and repeat until all checks pass. Does not post comments — this is a local developer tool.
Fix a GitHub issue in NeMo Speech (NVIDIA-NeMo/NeMo). Read the issue, reproduce the bug with a failing test, implement the fix, and verify tests pass. Only opens a PR if the user explicitly asks for it.
Run style checks and tests on changed files to verify code quality before committing.
Debug distributed training failures (NeMo, Megatron, PyTorch) from worker stderr logs and optional AIStore daemon logs. Finds root cause across NCCL timeouts, data loading errors, and storage failures.
| name | babysit-pr |
| description | Get a pull request to green CI. Diagnose and fix CI failures, push fixes, re-trigger CI via the "Run CICD" label, and repeat until all checks pass. Does not post comments — this is a local developer tool. |
Get a PR's CI to green. Nothing else — no review comments, no PR comments, no status summaries. This skill is run locally by developers in their own sandboxes.
The PR number is the primary input. It may come from:
/babysit-pr 567If no PR number is clear, ask for it before proceeding.
gh pr view <PR_NUMBER> --repo NVIDIA-NeMo/NeMo
gh pr checks <PR_NUMBER> --repo NVIDIA-NeMo/NeMo
gh pr diff <PR_NUMBER> --repo NVIDIA-NeMo/NeMo
Determine the current state:
| State | What to do |
|---|---|
| CI failing | Diagnose and fix (Step 2) |
| Merge conflicts | Resolve (Step 3) |
| Formatting check failing | Wait — do NOT fix (see below) |
| CI green | Done, nothing to do |
| CI pending | Wait for it to finish, then reassess |
The "Isort and Black Formatting" workflow (reformat_with_isort_and_black job) auto-pushes formatting fixes. If that check is failing or pending:
Check out the PR branch and inspect the failure logs:
gh pr checkout <PR_NUMBER> --repo NVIDIA-NeMo/NeMo
gh run list --repo NVIDIA-NeMo/NeMo --branch <branch-name>
gh run view <RUN_ID> --repo NVIDIA-NeMo/NeMo --log-failed
Before attempting a fix, check git log for recent commits. If you see a previous fix attempt that addressed the same failure and it is still failing, stop and tell the user — the issue needs human attention. Do not keep retrying the same fix.
Otherwise, identify the root cause, fix the code, and push:
git add <changed files>
git commit -s -m "<brief summary of fix>"
git push
After pushing a fix, add the "Run CICD" label to re-trigger the CI pipeline:
gh pr edit <PR_NUMBER> --repo NVIDIA-NeMo/NeMo --add-label "Run CICD"
The "CICD NeMo" workflow is triggered by this label and removes it automatically when done.
Then wait for CI to complete and reassess. Go back to Step 1.
If the PR branch has fallen behind main and has conflicts:
git fetch origin main
git rebase origin/main
# Resolve conflicts — keep the PR's intent, adopt refactors from main
git add <resolved files>
git rebase --continue
git push --force-with-lease
After rebasing, go back to Step 3 to re-trigger CI.
Isort and Black Formatting action handles that automatically.CI is green (or the only remaining failures are pre-existing / flaky and you've told the user about them). That's it.