AN

analyze-ci

It retrieves and parses failed CI logs to summarize specific errors and failing test cases.

Install

mkdir -p .claude/skills/analyze-ci && curl -L -o skill.zip "https://agentskills.codes/api/skills/download/2952" && unzip -o skill.zip -d .claude/skills/analyze-ci && rm skill.zip

Installs to .claude/skills/analyze-ci

Activation

This is the description your AI agent reads to decide when to run this skill — the better it matches your request, the more reliably it fires.

Analyze failed GitHub Action jobs for a pull request.
53 charsno explicit “when” trigger
Intermediate

Key capabilities

  • →Fetch logs from failed GitHub Action jobs
  • →Identify root causes and specific error messages
  • →Extract pytest test names from logs
  • →Generate per-job failure summaries
  • →Provide paths to raw and failed-step logs

How it works

The skill fetches logs using a package-specific command and then parses the failed-step log to identify root causes and error details.

Inputs & outputs

You give it
URL of a failed GitHub Action job, workflow run, or pull request
You get back
A focused summary containing workflow name, failed step, error context, and log paths

When to use analyze-ci

  • →Debug failing pull request CI jobs
  • →Summarize root causes of test failures
  • →Identify package version issues in builds

About this skill

Analyze CI Failures

Fetch logs from failed GitHub Action jobs and produce a focused per-job failure summary.

If no URLs are passed as arguments, use the relevant job, run, or PR URLs already shared in the conversation. If none are available, ask for a URL and stop.

Steps

  1. Fetch logs with the following command. Single-quote each URL to protect shell special characters like ?.

    uv run --package skills skills fetch-logs '<url1>' ['<url2>' ...]
    
  2. Read all reported ranges. Inspect the raw log directly if no failed-step range is available, then identify:

    • The root cause and its impact.
    • A short log excerpt showing the failure.
    • A full node ID for pytest failures (e.g. tests/test_foo.py::test_bar[case1]).
    • Package versions or snapshot changes that help explain the failure.
  3. Format each job with a heading, an evidence list, and a summary of 1-2 sentences plus the log excerpt.

    ## <workflow name> / <job name>
    
    - URL: <job_url>
    - Raw log: <raw_log_path>
    - <additional fields, if present>
    
    <summary>
    

    Copy the raw log path verbatim from step 1. If no failed step is recorded, include the reported Conclusion: field too.

When not to use it

  • →When the job URL is not for a GitHub Action

Prerequisites

GitHub Token

Limitations

  • →Requires specific URL formats for GitHub Action jobs
  • →Relies on the availability of cached log files

How it compares

Unlike manual log inspection, this skill automatically aggregates failure data and formats it into a structured summary for debugging.

Compared to similar skills

analyze-ci side by side with the closest alternatives in the catalog.

SkillInstallsUpdatedSafetyDifficulty
analyze-ci (this skill)16moReviewIntermediate
moai-workflow-testing15moReviewIntermediate
sentry-pr-code-review06moReviewIntermediate
python-testing-patterns776moReviewIntermediate

Try saying

Example prompts that trigger this skill in your AI assistant.

Search skills

Search the agent skills registry