PlugWiki

deployment-pipeline-design Skill (wshobson) for Claude Code

Helps Claude design multi-stage CI/CD pipelines with approval gates, canary or blue-green rollouts, health checks and rollback plans.

Installs
14K
on skills.sh
Stars
42.1K
on GitHub
Files
2
besides SKILL.md
Updated
Oct 2026
last repo commit

You describe your app, deploy target, environments and rollout needs. Claude then proposes pipeline stages, gate rules, a deployment strategy (rolling, blue-green, canary or feature flags), health checks and a rollback plan. SKILL.md lists the outputs as pipeline config, strategy, health checks, gates and a rollback plan.

The skill is advice and example config. The examples in its reference files use GitHub Actions, GitLab CI, Azure Pipelines, Argo Rollouts and kubectl. Claude only runs them if you ask it to, so review any command before it touches a cluster.

Good for

  • Platform and DevOps engineers designing a new release pipeline
  • Teams adding approval gates or canary rollouts to an existing pipeline
  • Developers debugging a pipeline whose stages pass but production misbehaves

Install deployment-pipeline-design

Claude Code

  1. 1

    Copy this prompt.

    Install the deployment-pipeline-design skill in Claude Code. Run this command:
    
    npx skills add wshobson/agents --skill deployment-pipeline-design
    
    Then check that it loaded and tell me how to use it.
  2. 2

    Paste it into Claude Code. It shows the change and asks before running it.

Requires Node.js.

The README says the skills routes install a skill without the plugin's agents, commands or hooks. Pick the plugin only if you want the whole cicd-automation set.

Built from the package data we hold. Not yet tested by us. Source.

What's in the skill

When it loads

Design multi-stage CI/CD pipelines with approval gates, security checks, and deployment orchestration. Use this skill when designing zero-downtime deployment pipelines, implementing canary rollout strategies, setting up multi-environment promotion workflows, or debugging failed deployment gates in CI/CD.

Sections in SKILL.md

  1. Purpose
  2. Input / Output
  3. What You Provide
  4. What This Skill Produces
  5. When to Use
  6. Detailed patterns and worked examples
  7. Troubleshooting
  8. Health check passes in pipeline but service is unhealthy in production
  9. Canary deployment never promotes to 100%
  10. Staging deploy succeeds but production job never starts
  11. Docker layer cache busted on every run causing slow builds
  12. Rollback leaves database migrations applied to old code
  13. Advanced Topics
  14. Related Skills

Key facts

Author
wshobson
Repository
wshobson/agents
Installs on skills.sh
14,035 as of 2026-11-02
Last repo commit
2026-10-30

What it includes

  • Stages, job order, parallelism and caching

    SKILL.md says it produces stage definitions, job dependencies, parallelism and a caching strategy from the app type, target and environment layout you give it.

    Try: “Design a staging-to-production pipeline for a Node service on Kubernetes with a manual approval before prod.”

    Pipeline design

  • Manual, time-based, multi-approver and metric gates

    details.md shows four gate patterns: manual approval in GitHub Actions, time-based in GitLab CI, multi-approver in Azure Pipelines and an automated Argo Rollouts metric gate.

    Approval gates

  • Rolling, blue-green, canary and feature flags

    A decision table and examples for four strategies. The blue-green and canary examples use kubectl and Argo Rollouts commands.

    Deployment strategies

  • Shallow versus deep readiness checks and a verify script

    Explains why a shallow ping can pass while the service is unhealthy, and includes a post-deployment verification script that calls a /health/ready endpoint with curl.

    Health checks

  • Automatic and manual rollback steps

    Covers kubectl rollout undo, rollback on failed checks and migration rollback with undo scripts.

    Rollback plan

  • Five common failures and fixes

    SKILL.md covers a health check that passes but the service is unhealthy, a canary that never promotes, a production job that never starts, a busted Docker layer cache and a rollback that leaves migrations applied.

    Troubleshooting

  • Platform configs, multi-region canary, migrations, freeze windows

    advanced-strategies.md has full pipelines for GitHub Actions, GitLab CI and Azure Pipelines, multi-region canary, expand/contract database migrations, a deployment freeze check script and a Slack notification script.

    Advanced strategies

From the official docs

Permissions

Medium risk

The skill bundles no scripts, but its examples include commands that can deploy to and change a Kubernetes cluster. Review each command before Claude runs it.

What it can do

  • Read your data
  • Create and change things
  • Delete things
  • Read-only mode

Watch out for

  • Examples include kubectl, docker and migration commands

    The references show kubectl apply, rollout undo, scale to zero, docker push and Flyway undo. Do not let an agent run them against production without reading them. github.com

  • One example drops a database column

    advanced-strategies.md shows an ALTER TABLE DROP COLUMN statement in the contract step of an expand/contract migration. SKILL.md says destructive migrations wait until the old code is fully retired. github.com

Security advisories

No published security advisories as of Nov 2, 2026.

Technical details
Turning on read-onlyNo read-only switch is documented. Ask Claude to propose config without running it.
Talks toExample commands call kubectl, docker, curl and cloud CLIs against your own cluster, registry and CI
Sign-in and keysThe examples reference CI secrets such as a Slack webhook and registry tokens. Those belong to your own CI system.
Scopes
  • Claude writes pipeline files and scripts in your project when you ask

  • The plugin route adds 5 agents and 1 command from the cicd-automation folder

Questions

What does the deployment-pipeline-design skill do?

It guides Claude through designing a CI/CD pipeline: stages, approval gates, a rollout strategy, health checks and rollback. You give it your app type, target, environments and rollout needs, and it produces config and a plan.

How do I install it?

Use the install box above. The skills CLI route installs only this skill. The Claude Code and Codex plugin routes install the whole cicd-automation plugin, which also adds agents and a command.

Does it deploy anything for me?

SKILL.md describes design output, not deployment. The reference files do contain kubectl, Argo Rollouts, Flyway and docker push commands as examples, so check any command Claude proposes before you run it against a real cluster or registry.

Does it need an API key or other setup?

The skill itself does not. The example pipelines refer to secrets such as a Slack webhook and registry or cloud credentials, which belong to your own CI system.

Which CI systems does it cover?

GitHub Actions, GitLab CI and Azure Pipelines have full examples in advanced-strategies.md. SKILL.md also points to the separate github-actions-templates and gitlab-ci-patterns skills for implementation detail, and secrets-management for secrets.

Is it safe to install?

It bundles no scripts. The risk comes from the example commands in its references, which can deploy, scale or delete Kubernetes resources and run database migrations if you or Claude execute them. Read each command first.

Information

updated Nov 2, 2026

Repository
wshobson/agents
Folder
plugins/cicd-automation/skills/deployment-pipeline-design

wshobson/agentsskills.sh

Sources

Docs, issue trackers and discussions this page is built from.

Show all 6
  1. Using skills in Claude · support.claude.com
  2. wshobson/agents README · github.com
  3. SKILL.md · github.com
  4. details.md · github.com
  5. advanced-strategies.md · github.com
  6. cicd-automation plugin.json · github.com

Updated Nov 2, 2026.