← ClaudeAtlas

ci-pipelinelisted

Design and troubleshoot CI/CD pipelines — GitHub Actions, GitLab CI, Jenkins. Use when setting up a CI pipeline, configuring automated testing and deployment, choosing a deployment strategy, or debugging a pipeline failure. This skill is specific to CI/CD pipeline design and does NOT cover Docker containerization, monitoring/alerting, or secret management.
ecoma-io/touchstone · ★ 1 · DevOps & Infrastructure · score 63
Install: claude install-skill ecoma-io/touchstone
# CI/CD Pipeline A CI/CD pipeline automates the path from commit to production. It answers: "does this change break anything, and how does it reach users?" ## Pipeline as code The pipeline definition lives in version control, not in a UI. `.github/workflows/`, `.gitlab-ci.yml`, `Jenkinsfile` — the config IS the pipeline. ## Fast feedback The fastest check that can catch a failure class should run first. Lint before type-check before unit test before integration test before deploy. A developer waiting 20 minutes for a lint error has already context-switched. ## Determinism The same commit run through the same pipeline twice must produce the same result. Cache keys must be content-addressed. External dependencies must be pinned. "Works on my machine" means the pipeline is missing a dependency. ## Secrets in CI Never echo a secret. Never pass a secret to a script that logs its arguments. Use the platform's secret store (GitHub Secrets, GitLab CI Variables) and inject at runtime, not at build time. ## Deployment strategy Blue-green, canary, or rolling — pick one and encode it. The pipeline must know how to roll back faster than it deployed. A deploy that can't be undone is a manual step wearing a YAML costume.