Automation & Ops ✓ NVIDIA · Official github-actionsci-triagebug-reportinggithub-clipytestmegatron-core

Megatron-Core CI Failure to Issue

Investigate failed GitHub Actions jobs and create structured bug issues for NVIDIA Megatron-LM.

FollowSkills review · FSRS-2.0
Not recommended
46/ 100 5-point scale 2.3 / 5
Trust10 / 25 · 2.0/5

The workflow is scoped to reading GitHub Actions data, inspecting PR/commit metadata, and creating an issue in NVIDIA/Megatron-LM. It requires duplicate checking, quoting actual logs, and stopping when a run is incomplete or logs are unavailable; no malware or credential theft is evident. However, issue creation and automatic assignment are external side effects without a mandatory final user confirmation, and least-privilege permissions, sensitive-log handling, rollback, and revocation are not specified, so points are deducted.

Reliability8 / 20 · 2.0/5

URL forms, failed-job selection, log extraction, PR/author resolution, duplicate checking, issue construction, and incomplete-run handling are described coherently. Reliability is reduced by dependence on gh, GitHub access, and network availability, plus limited handling and diagnostics for malformed inputs, pagination, redirected logs, and lookup failures. No skill-specific executable test evidence is supplied. Under static calibration the score remains at or below 10.

Adaptability9 / 15 · 3.0/5

The audience, accepted URL inputs, trigger phrases, and intended output are clear for NVIDIA/Megatron-LM CI triage. The skill is nevertheless tightly coupled to GitHub Actions and one repository, gives limited non-fit boundaries, and provides no Chinese-language or mainland-China reachability guidance; its core function depends on GitHub access, so points are deducted.

Convention8 / 15 · 2.7/5

The SKILL.md has readable metadata, progressive workflow steps, commands, an issue template, limitations, and failure guidance. The skill card supplies owner, Apache-2.0 terms, use case, risks, and a version; repository-level license, security, and publishing files add governance context. It lacks recommended Instructions, Examples, Purpose, and tags, plus a dedicated FAQ, changelog, and clear maintenance/update path; repository license metadata is also NOASSERTION, so it is not full marks.

Effectiveness7 / 15 · 2.3/5

The skill covers the core path from a failed CI run to a structured issue, including duplicate prevention, concise error capture, PR linkage, author assignment, and reporting the resulting URL. Its practical value is plausible, but no representative live output or third-party execution evidence is provided, GitHub permissions/log availability can block completion, and human review remains necessary. Static calibration limits this to 7.

Verifiability4 / 10 · 2.0/5

The commands, queried fields, and issue template are auditable, and BENCHMARK.md records static validation findings. However, Tier 3 details are unavailable, evals.json is empty, and there is no committed test suite or independent reproduction covering this skill's key paths. Evidence is therefore limited; the static score stays below the 5-point ceiling.

Evidence confidence:Low Reviewed Jul 20, 2026 Reviewed revision 55f18499943e
Before you use it
  • Issue creation and automatic assignment have external effects and may require elevated permissions; add explicit user confirmation and verify the repository, title, body, labels, and assignee before execution.
  • GitHub Actions logs may contain tokens, internal paths, or other sensitive data; the skill does not define redaction, least-privilege access, or retention handling.
  • The core workflow depends on gh, GitHub API access, repository permissions, and network reachability; no Chinese-language guidance or mainland-China fallback is provided.
  • Tier 3 evaluation details are unavailable and evals.json is empty; do not treat the benchmark's PASS verdict as completed real end-to-end validation.
See the full review method →

What does this skill do, and when should you use it?

This skill handles failed GitHub Actions runs or jobs for NVIDIA Megatron-LM. It parses a run or job URL, identifies failed jobs, extracts failure details from logs, and resolves the Pull Request that triggered the run. It also finds the latest author of the failing test file, checks for duplicate open issues, and creates and assigns a new bug issue only when no matching issue exists. It is intended for engineering teams that need to turn CI failures into trackable work.

Reads a GitHub Actions run or job URL; uses the GitHub CLI to inspect runs, jobs, and failed jobs; retrieves job logs through the GitHub API; filters lines containing FAILED, ERROR, assertions, Traceback, or Exception; extracts failed test node IDs, job names, run and job URLs, and Pull Request details; queries the latest contributor to the failing test file; searches for open duplicate issues; when no duplicate is found, creates a bug-labeled issue in NVIDIA/Megatron-LM containing the triggering PR, error summary, and reproduction command, and attempts to assign it to the test file author; finally reports the new or duplicate issue URL.

  1. A developer shares a failed GitHub Actions job URL and wants a Megatron-LM bug report created.
  2. A maintainer shares only a failed run URL and needs one or all failed jobs triaged.
  3. A test maintainer wants the CI issue assigned to the latest author of the failing test file.
  4. An engineering team wants duplicate open issues checked before filing a new CI defect.

What are this skill's strengths and limitations?

Pros
  • Covers the workflow from log investigation and error extraction through author lookup and issue creation.
  • Requires a duplicate-issue check before creating a new report.
  • Uses an issue structure containing the failing test, run, job, PR, error, and reproduction details.
  • Keeps the reported error excerpt concise at roughly 30 lines.
Limitations
  • The workflow is specifically fixed to NVIDIA/Megatron-LM CI failures and issues.
  • It depends on the GitHub CLI, GitHub API access, and permissions to create and assign issues.
  • The source provides no automated test suite or evidence of real-platform validation.
  • If the run is still in progress or logs are unavailable, it can only ask the user to retry later.

How do you install this skill?

Install the specific skill with the default skills CLI: npx skills add nvidia/skills --skill mcore-create-issue --yes. The README does not document additional skill-specific installation steps.

How do you use this skill?

Give a loaded Agent a GitHub Actions run or job URL and use a request such as “create an issue for this failure” or “file a bug for this CI run.” Supported formats are https://github.com/<owner>/<repo>/actions/runs/<run_id> and the job form with /job/<job_id>. If a run has multiple failed jobs, select the job to triage or explicitly request that all of them be triaged.

FAQ

Can it handle Actions failures from any GitHub repository?
That is not established by the source. Although the input URL may reference another repository, the workflow commands are fixed to NVIDIA/Megatron-LM for jobs, pull requests, logs, and issues.
What happens when a duplicate issue already exists?
The skill must stop without creating a new issue and return the existing issue link.
Does it infer the root cause automatically?
No. The instructions require quoting actual log output and prohibit guessing the root cause.
Is installation or use paid?
The supplied materials do not specify any cost.

More skills from this repository

All from NVIDIA/skills

Dev & Engineering ✓ NVIDIA · Official

Megatron-Core Testing Guide

Design, run, and reproduce distributed Megatron-Core tests with CI-aligned workflows.

Dev & Engineering ✓ NVIDIA · Official

CODEOWNERS PR Splitter

Split large pull requests by CODEOWNERS groups to reduce review burden while keeping each PR independently mergeable.

Dev & Engineering ✓ NVIDIA · Official

Megatron-Bridge Communication Overlap Guide

Configure and verify TP, DP, and PP communication overlap in Megatron-Bridge training.

Dev & Engineering ✓ NVIDIA · Official

MoE Hardware Configuration Reference

Plan MoE training layouts and throughput expectations across NVIDIA GPU platforms.

Dev & Engineering ✓ NVIDIA · Official

Megatron-Core Linting & Formatting

Standardize Megatron-LM Python checks and formatting before code review.

Dev & Engineering ✓ NVIDIA · Official

NeMo MBridge Hierarchical Context Parallelism

Enable, troubleshoot, and verify hierarchical context parallelism in Megatron-Bridge.

Dev & Engineering ✓ NVIDIA · Official

TileGym cuTile Kernel Guide

A step-by-step guide for registering, implementing, testing, and benchmarking a new cuTile GPU kernel in TileGym.

Dev & Engineering ✓ NVIDIA · Official

NVIDIA cuOpt Developer

Guides source-level cuOpt modification, builds, tests, debugging, and contributions.

Dev & Engineering ✓ NVIDIA · Official

NeMo Relay Installation Guide

Choose and verify the right NeMo Relay path for CLIs, language packages, and maintained frameworks.

Dev & Engineering ✓ NVIDIA · Official

NeMo MBridge GPU Memory Tuning

Diagnose GPU OOMs in Megatron Bridge training and reduce fragmentation, activation, and PEFT memory usage.

Dev & Engineering ✓ NVIDIA · Official

NeMo Relay Quick Start

Helps first-time NeMo Relay users prove observable execution value through the smallest suitable trial.

Dev & Engineering ✓ NVIDIA · Official

Megatron Bridge Activation Recompute

Trade extra compute for lower GPU activation memory in Megatron Bridge training.

Dev & Engineering ✓ NVIDIA · Official

NeMo Evaluator Plugin Skill

Run and manage NeMo Platform model evaluations through the Evaluator CLI and Python SDK.

Dev & Engineering ✓ NVIDIA · Official

NeMo MBridge Parallelism Strategy Guide

Choose, combine, and validate Megatron Bridge parallelism configurations.

Dev & Engineering ✓ NVIDIA · Official

NeMo MBridge MoE Expert-Parallel Overlap

Configure, validate, and troubleshoot MoE expert-parallel communication overlap in Megatron-Bridge.

Dev & Engineering ✓ NVIDIA · Official

NeMo MBridge MoE VLM Training Guide

Guides FSDP and 3D-parallel choices and tuning for MoE vision-language model training.

Dev & Engineering ✓ NVIDIA · Official

NeMo MBridge Training Resiliency

Configure fault recovery and hang detection for Megatron Bridge distributed training.

Dev & Engineering ✓ NVIDIA · Official

NeMo Relay Migration Assistant

Safely migrate NeMo Flow projects to NeMo Relay.

Automation & Ops ✓ NVIDIA · Official

DOCA Telemetry Utils

Discover telemetry counters, translate Data IDs, and verify BlueField support before committing exporter configuration.

Automation & Ops ✓ NVIDIA · Official

Megatron-LM on SLURM

Launch, monitor, and troubleshoot multi-node Megatron-LM training on SLURM GPU clusters.

Related skills