‹ 首页

ci-cd-pipelines

@developersglobal · 收录于 1 周前

Automated quality gates from commit to production. Every merge to main is potentially shippable. No manual steps in the deployment path.

场景代码合并上线发布

适合你，如果每次合并到主分支都想自动部署到生产环境

/ 下载安装
ci-cd-pipelines.skill双击,或拖进 Claude 桌面版 / Cowork,即完成安装↓ .skill↓ .zip
用别的 agent?下载 .zip 解压,把文件夹放进它的技能目录
Claude Code~/.claude/skills/(项目级 .claude/skills/)
Codex CLI~/.codex/skills/
Cursor自动读取上面两处目录
其他工具见其文档的「skills」目录;两个下载是同一份文件,只是名字不同
/ 通过 npx 安装 校验哈希
npx oh-my-skill add developersglobal/ai-agent-skills/ci-cd-pipelines
/ 通过 bash 安装
curl -fsSL https://oh-my-skill.com/install.sh | bash -s -- developersglobal/ai-agent-skills/ci-cd-pipelines
/ 已经装过?验证本机副本,不用重装
npx oh-my-skill verify developersglobal/ai-agent-skills/ci-cd-pipelines
安装目标可用 --agent / --scope 或 --to 明确指定;省略时只会在唯一已存在的 agent 目录上自动选择,零命中或多命中会停止并提示。content_hash 缺失或不一致均拒装。
64GitHub stars
~634上下文体积 · 单文件
镜像托管

怎么用

技能原文 SKILL.md作者撰写 · MIT · f47f948
Overview

CI/CD is the automation layer that enforces quality gates consistently, without relying on human memory or discipline. When CI is green, you know the code is tested, linted, and deployable. When it's red, nothing ships.

When to Use
  • Setting up a new project
  • Adding a new quality gate
  • Reviewing CI/CD pipeline configuration
Process
Step 1: CI Gates (Every PR)
  1. All gates must pass before merge is allowed:
  2. Lint: code style and static analysis
  3. Unit tests: all pass
  4. Integration tests: key boundaries covered
  5. Security scan: SAST, dependency vulnerabilities
  6. Build: production artifact builds successfully
  7. Gates run in parallel where possible (speed matters).
  8. Maximum CI time: 10 minutes. If slower, optimize.

Verify: Merging is blocked when any gate fails.

Step 2: CD Pipeline (Every Main Merge)
  1. Main branch is always deployable.
  2. Deployment pipeline:
  3. Deploy to staging → run smoke tests → deploy to production (canary) → full rollout
  4. Every step is automated — no manual "click to deploy."
  5. Rollback is automated and tested.

Verify: A push to main triggers automated deployment with no human intervention required.

Step 3: Feature Flags Over Feature Branches
  1. Incomplete features go behind feature flags — not long-lived branches.
  2. Feature flags allow dark launching, A/B testing, and instant rollback without redeployment.
  3. Feature flag state is tracked in a dashboard.

Verify: New features are behind flags. No feature branches > 2 days old.

Step 4: Pipeline as Code
  1. CI/CD config is in the repo (.github/workflows/, .gitlab-ci.yml, etc.).
  2. Pipeline changes go through code review like any other change.
  3. Pipeline config is tested: changes to CI don't break CI.

Verify: CI config is in the repo and reviewed.

Common Rationalizations (and Rebuttals)

| Excuse | Rebuttal | |--------|----------| | "It's a small change, CI is optional" | Every "small change" that skipped CI is in the origin story of a major incident. | | "Manual deployment gives us control" | Manual steps introduce human error. Automation gives you control. | | "CI is too slow" | Optimize it. Don't skip it. |

Verification
  • [ ] All quality gates run on every PR
  • [ ] Merge blocked when any gate fails
  • [ ] Main → production deployment is fully automated
  • [ ] Rollback is automated and tested
  • [ ] Feature flags in place for incomplete features
References
  • [production-deployment skill](../production-deployment/SKILL.md)
  • [git-workflow skill](../git-workflow/SKILL.md)
按 MIT 许可原样转载,未经改动 · 在 GitHub 查看 →

评论

登录即可评论;带「已验证安装」的,是发布者名下有本店的安装或持有记录。
…