CO

country-atlas-ci-fix

Automates the identification and resolution of CI/CD pipeline failures using targeted reproduction and patches.

Install

mkdir -p .claude/skills/country-atlas-ci-fix && curl -L -o skill.zip "https://agentskills.codes/api/skills/download/18895" && unzip -o skill.zip -d .claude/skills/country-atlas-ci-fix && rm skill.zip

Installs to .claude/skills/country-atlas-ci-fix

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.

Use when GitHub Actions, pre-commit, tests, typing, lint, SQL lint, Docker smoke, or Playwright checks fail.
108 chars✓ has a “when” trigger
Intermediate

Key capabilities

  • Identify the failing job, command, file, line, and commit from CI logs
  • Reproduce failing commands locally when possible
  • Apply minimal, behavior-preserving patches to fix root causes
  • Run targeted failing checks after applying fixes
  • Validate fixes with relevant quality gates

How it works

The skill analyzes CI failure logs to pinpoint the exact failing command and context, then attempts to reproduce the issue locally, applies a minimal fix, and validates it by re-running the targeted checks and quality gates.

Inputs & outputs

You give it
Exact failing command and log excerpt from a CI failure
You get back
A minimal, behavior-preserving patch and validation results from targeted checks and quality gates

When to use country-atlas-ci-fix

  • Fixing failing GitHub Actions
  • Resolving lint or type-check failures
  • Debugging Docker smoke tests

About this skill

CI Fix Workflow

Start from the exact failing command and log excerpt.

Workflow

  1. Read AGENTS.md.
  2. Capture the failing job, command, file, line, and commit.
  3. Reproduce the smallest failing command locally when possible.
  4. Fix the root cause with the smallest behavior-preserving patch.
  5. Run the targeted failing check.
  6. Run the relevant quality gate when feasible:
python dev_tools_scripts_runner.py --profile quick

Notes

  • If gh or GitHub connector access is unavailable, say so and continue with local reproduction.
  • Prefer deterministic test fixes over loosening assertions.
  • A monkeypatch that stops intercepting a call after a module split is a known project-specific failure mode (see .ai/project/12-domain-rules.md).
  • Do not push to main unless the current user turn explicitly asks for it.

When not to use it

  • When the user explicitly asks to push to main without prior validation
  • When `gh` or GitHub connector access is unavailable and local reproduction is not feasible
  • When the issue is not related to GitHub Actions, pre-commit, tests, typing, lint, SQL lint, Docker smoke, or Playwright checks

Limitations

  • Does not push to `main` unless explicitly asked by the user
  • Requires the exact failing command and log excerpt as input
  • Prefers deterministic test fixes over loosening assertions

How it compares

This skill automates the diagnosis and patching of CI failures with validation, contrasting with manual debugging and iterative testing.

Compared to similar skills

country-atlas-ci-fix side by side with the closest alternatives in the catalog.

SkillInstallsUpdatedSafetyDifficulty
country-atlas-ci-fix (this skill)022dNo flagsIntermediate
dev25moReviewAdvanced
testing01moReviewAdvanced
ci-pr-helper06moReviewBeginner

Try saying

Example prompts that trigger this skill in your AI assistant.

Search skills

Search the agent skills registry