gh-ci-investigate

Diagnose GitHub Actions CI failures by collecting failed job logs via the gh CLI.

Updated Jun 24, 2026
One-click install
npx skills add https://github.com/Hakkadaikon/hymme --skill gh-ci-investigate-hakkadaikon
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: gh-ci-investigate
Source: https://github.com/Hakkadaikon/hymme/tree/main/skills/gh-ci-investigate
Command: npx skills add https://github.com/Hakkadaikon/hymme --skill gh-ci-investigate-hakkadaikon

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve? Investigating a failed GitHub Actions run normally means clicking through the web UI to find which job and step failed and then reading the logs. WebFetch cannot retrieve these logs (it only returns HTML), so this Skill uses the gh CLI through a deterministic script to reliably identify the failed workflow, job, and step, and extract only the failing step's logs. ## Core Features & Use Cases - Input resolution: Accepts run/PR/job URLs, run IDs, PR numbers, or defaults to the latest run on the current branch, resolving them to a concrete target. - Preflight summary: Reports authentication status, failed jobs and steps, and their conclusions (failure, cancelled, timed_out) in a structured digest. - Targeted log extraction: Uses gh run view --log-failed to fetch only the failing step's logs, avoiding context bloat from full logs. - Use Case: A teammate pastes a failing CI run URL and asks why the build broke. The Skill identifies the failed job and step, pulls the relevant log lines, cross-references the repository code, and proposes a fix in systematic-debugging Phase 4 format without applying changes automatically. ## Quick Start Paste the failing GitHub Actions run URL and ask the assistant to investigate why the CI failed and propose a fix.

Frequently Asked Questions about gh-ci-investigate

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I investigate a failed GitHub Actions run from the command line?

Use the gh CLI with `gh run view --log-failed` to retrieve only the failing step's logs. This Skill wraps that in a script that accepts a run URL, run ID, or PR number and summarizes the failed jobs and steps first.

How to get logs for a failed CI job on GitHub?

Run the logs subcommand with the run URL or ID, optionally adding --job with a job ID to narrow scope. It calls `gh run view --log-failed`, which returns only failed step logs instead of the full log output.

Can WebFetch read GitHub Actions logs?

No. WebFetch against github.com only returns the Actions page HTML, not the actual log content. Logs must be retrieved through the gh CLI or the GitHub API, which is why this Skill routes all collection through gh.

Does this work with GitLab CI or other CI platforms?

No. The Skill targets GitHub Actions on github.com and GitHub Enterprise only, since it relies on the gh CLI. For GitLab CI or other platforms, use the corresponding platform CLI instead.

Why does CI log retrieval fail with authentication errors?

Private repositories require a gh token with read permissions. Check `gh auth status`, and for GitHub Enterprise run `gh auth login --hostname <host>`. The Skill reports auth status in preflight and stops rather than guessing.

Does the gh-ci-investigate skill apply fixes automatically?

No. It stops at identifying the root cause with log evidence and presenting a fix proposal in systematic-debugging Phase 4 format. Actual code or workflow changes require user approval and follow the normal editing flow.