Index of every page on this site: llms.txt
All skills
nvidia avatar

/mcore-create-issue

@6b75e9e
by NVIDIA Corporationnvidia/skills4.1k stars
547
Run it yourself

Investigate a failing GitHub Actions run or job and create a GitHub issue for the failure.

  • 5 files
  • 15.9 KB
  • Apache-2
  • Updated 6 months ago
  • GitHub
Use this Skill: https://skilld.dev/gh/nvidia/skills/mcore-create-issue

Nothing lands on disk. Nothing to clean up.

Fork this Skill

Edit a local copy. It keeps the author and licence.

SKILL.md

≈28 tokens for metadata: the name and description. ≈1.5k when used: this file. ≈1.3k more on demand in 3 files.

Description uses 3.5% of example budget

Before choosing a Skill, your Agent reads its name and description. All available Skills share that space.

  • A shorter description leaves more room for other Skills. This entry exceeds our 1% size suggestion.

In our Claude Code example, all Skill names and descriptions share 8,000 characters. This Skill uses ≈280 characters, or 3.5%.

The 1% threshold is a size suggestion. Longer descriptions can still fit.

Your model, settings, and other Skills decide how much text your Agent can read.

Example settings and source

The example uses a 200k-token context and default Claude Code settings. The count includes the name, description, separators, and when_to_use when present. Codex also counts local file paths.

Skit's source and limits: Codex 0.160.1, Claude Code 2.1.292.

Triage CI Failure into a GitHub Issue

Investigate a failing GitHub Actions job, extract the root cause, and file a well-structured bug issue against NVIDIA/Megatron-LM.

Workflow

1. Parse the URL

The argument is a GitHub Actions URL. It will be one of:

  • Job URL: https://github.com/<owner>/<repo>/actions/runs/<run_id>/job/<job_id>
  • Run URL: https://github.com/<owner>/<repo>/actions/runs/<run_id>

Extract run_id and, if present, job_id.

2. Identify failed jobs

  • If a job_id was provided, use that job directly.

  • If only a run_id was provided, list all failed jobs in the run:

    gh run view <run_id> --repo NVIDIA/Megatron-LM --json jobs \
      --jq '[.jobs[] | select(.conclusion == "failure") | {id: .databaseId, name: .name, url: .url}]'

    If multiple jobs failed, ask the user which one to triage, or triage all of them if they say so.

3. Fetch the failure logs

For each failed job, retrieve the logs and narrow them down to the failure:

# Pull the raw log and keep only error-bearing lines
gh api repos/NVIDIA/Megatron-LM/actions/jobs/<job_id>/logs 2>&1 \
  | grep -E "(FAILED|ERROR|\bError\b|assert|Traceback|Exception|##\[error\])" \
  | head -200

Also capture the full job name:

gh run view --job <job_id> --repo NVIDIA/Megatron-LM --json name --jq .name

If the grep output is sparse, download the full logs and look for the pytest FAILURES section or the last non-zero exit signal.

4. Resolve the triggering PR and test author

Triggering PR: the run's head branch follows the pattern pull-request/<number>. Extract it and resolve the PR:

gh run view <run_id> --repo NVIDIA/Megatron-LM --json headBranch --jq .headBranch
# → e.g. "pull-request/4332"
# Extract PR number and fetch metadata:
gh pr view <pr_number> --repo NVIDIA/Megatron-LM --json number,title,url \
  --jq '{number: .number, title: .title, url: .url}'

Test file author: find the GitHub login of whoever last touched the failing test file. The file may not exist on main — first determine the PR's base branch, then search from there:

# 1. Get the PR's base branch (e.g. "main", "dev", "release/X.Y")
gh pr view <pr_number> --repo NVIDIA/Megatron-LM --json baseRefName --jq .baseRefName

# 2. Search commits on that base branch
gh api "repos/NVIDIA/Megatron-LM/commits?path=<test-file-path>&sha=<base-branch>&per_page=1" \
  --jq '.[0] | {login: .author.login, name: .commit.author.name, sha: .sha}'

If the result is empty (file was introduced by the PR itself), query the PR's commits instead:

gh api "repos/NVIDIA/Megatron-LM/pulls/<pr_number>/commits" \
  --jq '[.[] | select(.files? // [] | any(.filename == "<test-file-path>"))] | .[0].author.login'

As a last resort, list the PR commits and pick the author of the commit whose message most closely relates to the failing test file.

5. Extract the root cause

From the logs, identify:

  • Failed test(s): lines matching FAILED tests/...::... give the exact pytest node IDs.
  • Error message: the assertion failure, exception type, or first meaningful traceback frame — keep it under ~30 lines.
  • Job name: the GitHub Actions job name (e.g. tests/unit_tests/transformer/moe/**/*.py - latest).
  • Run / job URLs and PR URL: for linking in the issue.

6. Check for duplicate issues

Search for open issues that already cover the same test:

gh issue list --repo NVIDIA/Megatron-LM \
  --state open \
  --search "<failed-test-filename>" \
  --json number,title,url \
  --limit 10
  • If a matching open issue exists, do not create a new one. Report the existing issue to the user and stop.
  • If no match is found, proceed to file a new issue.

7. Create the issue

Pass --assignee <test-author-login> to assign the issue to the test file's author. Include the triggering PR URL in the issue body.

gh issue create \
  --repo NVIDIA/Megatron-LM \
  --title "🐛 CI failure: <failed-test-node-id>" \
  --label "bug" \
  --assignee "<test-author-login>" \
  --body "..."

Use the bug-report template body structure:

**Describe the bug**

CI test `<failed-test-node-id>` failed in job [`<job-name>`](<job-url>).
Tag @NVIDIA/mcore-oncall to get oncall's attention to this issue.

**Failing run**

| Field | Value |
|-------|-------|
| PR    | [#<pr_number>: <pr_title>](<pr_url>) |
| Run   | [<run_id>](<run_url>) |
| Job   | [<job_name>](<job_url>) |

**Error**

<core error message / traceback — 30 lines max>


**Steps/Code to reproduce bug**

Re-run the failing CI job linked above, or locally inside the dev container:

```bash
pytest <failed-test-node-id>

Additional context

Triaged automatically via /triage-issue.


If multiple tests failed in the same job, list each one as a separate bullet
under "Describe the bug" and include the combined error snippets. Assign the
issue to the author of whichever test file appears first in the failure list.

### 8. Report back to the user

Print the URL of the newly created issue (or the duplicate, if found) so the
user can review or share it.

## Important guidelines

- Never create an issue if a duplicate already exists — link the existing one instead.
- Always include the triggering PR link in the issue body.
- Always assign the issue to the test file's most recent author. If the author
  lookup fails (e.g. the commit was made by a bot or the login is unavailable),
  skip `--assignee` and note it in the "Additional context" section.
- Keep the error snippet concise (≤30 lines). Truncate long tracebacks and note that the full log is available via the job URL.
- Do not guess the root cause — quote the actual log output verbatim.
- If the job is still in progress or the logs are unavailable, say so and ask the user to retry once the run completes.

Source: SKILL.md on GitHub

skilld matched fixed text patterns in SKILL.md and file names. Patterns miss obfuscated code.

skilld run checks every file with the same patterns. It asks for approval before it loads a Skill with a behavior marked Needs approval.

No third-party reports yet.

Signed by skilld at 6b75e9e. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub yesterday.

Activeupdated 6 months ago
user_invocable
true
argument
GitHub Actions run or job URL
metadata
{
  "author": "Philip Petrakian <ppetrakian@nvidia.com>"
}
Other metadata
when_to_use
User shares a GitHub Actions URL and wants to file a bug report; 'create an issue for this failure', 'file a bug for this CI run', 'triage this GitHub Actions failure'.

README badge

README badge for nvidia/skills/mcore-create-issue