/deploy

/deploy

Route to the right deployment pattern based on your target environment and strategy.

Usage

/deploy                 — interactive deployment strategy guide
/deploy k8s             — Kubernetes deployment (blue-green, canary, rolling)
/deploy serverless      — serverless deployment (Lambda, Cloud Run, Vercel)
/deploy docker          — Docker Compose or Swarm deployment
/deploy rollback        — roll back a failed deployment

Steps Claude Should Follow

1. Identify Target and Environment

If not specified via $ARGUMENTS, ask:

• Deployment target: Kubernetes, serverless (Lambda/Cloud Run/Vercel), Docker Compose, bare metal?
• Target environment: dev / staging / production?

2. Load the Right Skill

Load skill deployment-patterns for strategy selection and execution steps. For pipeline setup, additionally load skill ci-cd-patterns.

3. Recommend Strategy by Target

Target	Environment	Recommended Strategy
Kubernetes	production	Argo Rollouts canary (5% → 25% → 100%)
Kubernetes	staging	Rolling update
Serverless	production	Traffic splitting via weighted aliases/revisions
Serverless	staging	Direct deploy
Docker	production	Blue-green with nginx upstream switch
Docker	staging	Recreate or rolling update

4. Pre-Deploy Readiness Check

Before executing, verify all of the following are in place:

• Readiness and liveness probes defined
• Rollback procedure documented (and tested)
• Monitoring alerts active for error rate and latency
• Environment variables validated at startup (fail-fast on misconfiguration)
• Health check endpoint responds correctly in the new build

If any item is missing, surface it and offer to add it before proceeding.

5. Execute or Guide

For automated targets: run the deployment commands step by step, showing output. For manual targets: provide exact commands with explanations.

6. Post-Deploy Verification

After deployment completes:

1. Poll health endpoint until stable (or surface failures immediately)
2. Show rollback command in case a revert is needed
3. Recommend next steps from the list below

Rollback (/deploy rollback)

1. Ask: which deployment to roll back (target, environment, previous version)?
2. Load skill deployment-patterns — Rollback Strategies section
3. Execute rollback for the identified target (kubectl rollout undo / previous Lambda version / nginx upstream swap)
4. Verify health after rollback

After This

• /resilience-review — validate health checks and failure modes post-deploy
• /add-observability — ensure deploy metrics and alerts are in place
• /slo — define or update error budget after a new deployment