Deploy Execute Skill

Execute deployment to target environment with safety controls.

Overview

Trigger: /popkit-ops:deploy execute or after validation passes

Purpose: Execute the actual deployment to the target environment. Supports dry-run mode for safe testing, captures deployment metrics and logs, and handles automatic rollback on failure.

Critical Rules

1. ALWAYS run validation before production deploys - Skip only for dry-run
2. ALWAYS confirm before live deployment - Require explicit user confirmation
3. ALWAYS support dry-run mode - Default to dry-run for first deployment
4. NEVER deploy with blocking validation failures - Respect go/no-go
5. ALWAYS record deployment in history - Update deploy.json after every attempt
6. ALWAYS capture rollback info - Save previous version for potential rollback

Required Decision Points

Step	When	Decision ID
1	Target selection	deploy_target
2	Mode selection	deploy_mode
3	Before live deploy	confirm_deploy
4	After deployment	post_deploy

Skipping these violates PopKit UX standard.

Process

Step 1: Prepare Deployment

python scripts/execute_deployment.py --dir . --action prepare --target docker

This:

• Loads deploy.json configuration
• Identifies target and environment
• Checks last deployment for rollback reference
• Prepares deployment plan

Step 2: Target Selection (MANDATORY)

Use AskUserQuestion:
- question: "Which target should I deploy?"
- header: "Target"
- options:
  - "Primary target" - Deploy to primary configured target
  - "All targets" - Deploy to all enabled targets
  - "Choose target" - Select specific target

Step 3: Mode Selection (MANDATORY)

Use AskUserQuestion:
- question: "How should I deploy?"
- header: "Mode"
- options:
  - "Dry run (Recommended first)" - Simulate without executing
  - "Deploy" - Execute the deployment
  - "Canary deploy" - Deploy to small percentage first

Step 4: Confirmation for Live Deploy (MANDATORY)

Use AskUserQuestion:
- question: "Ready to deploy? This will push changes to the target environment."
- header: "Confirm"
- options:
  - "Deploy now" - Execute
  - "Dry run first" - Simulate first
  - "Abort" - Cancel

Step 5: Execute

python scripts/execute_deployment.py --dir . --action execute --target docker --mode deploy --version 1.2.0

Step 6: Post-Deploy (MANDATORY)

Use AskUserQuestion:
- question: "Deployment complete. What next?"
- header: "Result"
- options:
  - "Verify deployment" - Health checks
  - "Rollback" - Something went wrong
  - "Done" - Deployment successful

Deployment Modes

Dry Run

• Simulates all deployment steps without executing
• Validates commands, configs, and access
• Reports what would happen
• No changes to target environment

Standard Deploy

• Full deployment execution
• Captures start/end timestamps
• Records deployment in history
• Triggers post-deploy verification

Canary Deploy

• Deploys to small subset (e.g., 10% of instances)
• Monitors for errors during canary period
• Promotes to full deployment or rolls back
• Only available for Kubernetes/Docker Swarm targets

Output Format

Deployment Execution
====================
Target: docker (primary)
Mode: deploy
Version: 1.2.0
Environment: production

Steps:
  [1/5] Building image... done (45s)
  [2/5] Pushing to registry... done (12s)
  [3/5] Updating deployment... done (3s)
  [4/5] Waiting for rollout... done (30s)
  [5/5] Health check... passed

Result: SUCCESS
Duration: 90s
Deployment ID: deploy-20260318-143022
Image: ghcr.io/myorg/myapp:1.2.0

History updated in .claude/popkit/deploy.json
Next: Verify with /popkit-ops:deploy validate or rollback with /popkit-ops:deploy rollback

Failure Handling

When deployment fails:

1. Capture error details - Full error message and stack trace
2. Determine failure point - Which step failed
3. Auto-rollback if configured - Roll back to previous version
4. Record failure in history - Status: "failed" with error details
5. Present options to user - Fix and retry, rollback, or abort

Integration

Called by:

• /popkit-ops:deploy execute command
• pop-deploy-validate workflow (proceed to deploy step)

Triggers:

• deployment-validator agent (post-deploy verification)

Followed by:

• pop-deploy-rollback - If deployment fails or needs rollback
• pop-deploy-validate - Post-deploy verification

Related

Component	Relationship
pop-deploy-validate	Pre-deploy validation
pop-deploy-rollback	Rollback on failure
deployment-validator	Agent for verification
deploy.json	Configuration and history