Deploy

Purpose

Deploy services to target environments via CI/CD workflow. Handles environment selection, deployment locking to prevent concurrent deploys, and triggers deployment with appropriate options.

When to Use

• •Deploying changes to staging or production
• •Releasing new features after PR merge
• •Hotfix deployments
• •Rollback to previous version

Quick Reference

• •Setup: /deploy configure (run once during framework setup)
• •Usage: /deploy (uses saved config)
• •Update: /deploy learn (re-analyze deployment setup)
• •Config: .claude/skills/deploy.yaml

Modes

Deployment Flow

┌─────────────────────────────────────────────────────────────────────┐
│  1. SELECT ENV  →  2. CHECK LOCK  →  3. DEPLOY  →  4. VERIFY        │
└─────────────────────────────────────────────────────────────────────┘

Step 1: Select Environment

Ask the user which environment to deploy to (configured via configure).

Step 2: Check Deployment Lock

Before deploying, check if another deployment is in progress:

# Example: GitHub Actions
gh run list --workflow=deploy.yaml --status=in_progress --json databaseId,status

# Example: Check for queued
gh run list --workflow=deploy.yaml --status=queued --json databaseId,status

Lock Rules:

• •If deployment is in_progress → STOP, wait for completion
• •If deployment is queued → STOP, wait or cancel
• •If no active deployments → PROCEED

Step 3: Trigger Deployment

Use the configured DEPLOY_COMMAND from .claude/skills/deploy.yaml:

# Examples based on platform:
# GitHub Actions
gh workflow run deploy.yaml -f environment=staging

# Kubernetes
kubectl apply -f k8s/staging/

# Task runner
task deploy:staging

# npm script
npm run deploy -- --env staging

Step 4: Monitor and Verify

# Watch workflow (GitHub Actions)
gh run watch

# Check status
gh run list --workflow=deploy.yaml --limit=1

# After completion, verify deployment
# (Use deploy-verify skill)

Configuration

Config Location

Config path depends on how the plugin was installed:

Precedence when reading (first found wins):

1. •.claude/skills/deploy.local.yaml
2. •.claude/skills/deploy.yaml
3. •Skill defaults

Config Fields

Configured via wizard during /deploy configure:

Deployment Scenarios

Standard Deployment (Most Common)

# Using configured command
${DEPLOY_COMMAND} -f environment=staging

Force Rebuild

When you need to rebuild (e.g., dependency updates):

${DEPLOY_COMMAND} -f environment=staging -f force_build=true

Deploy Specific Version

Rollback or deploy specific commit:

${DEPLOY_COMMAND} -f environment=staging -f version=<commit-sha>

Lock Management

Check Lock Status

# GitHub Actions example
IN_PROGRESS=$(gh run list --workflow=deploy.yaml --status=in_progress --json databaseId | jq length)
QUEUED=$(gh run list --workflow=deploy.yaml --status=queued --json databaseId | jq length)

if [ "$IN_PROGRESS" -gt 0 ] || [ "$QUEUED" -gt 0 ]; then
  echo "Deployment locked: $IN_PROGRESS in progress, $QUEUED queued"
else
  echo "No active deployments - safe to deploy"
fi

Cancel Stuck Deployment

If a deployment is stuck (use with caution):

# Get the run ID
gh run list --workflow=deploy.yaml --status=in_progress --json databaseId

# Cancel the run
gh run cancel <run-id>

Post-Deployment

After deployment completes:

1. •Verify with deploy-verify skill: /deploy-verify
2. •Check service health
3. •Monitor logs for errors

Automation

See skill.yaml for the full deployment procedure. See sharp-edges.yaml for common deployment pitfalls.

Related Skills