Implement and Deploy

Overview

Full lifecycle: implement feature, containerize, CI/CD pipeline, deploy to Azure, verify live.

Core principle: Modular phases with skip-if-done checks and user prompts at decision points.

Announce at start: "I'm using the implement-and-deploy skill to build and deploy this feature."

Phase 1: Repo Setup

Skip if: Already in a git repo with remote configured.

1. Ask user: create new repo or use existing?
2. If new: gh repo create <name> --public --clone and scaffold project
3. If existing: verify git remote -v has a GitHub remote
4. Create feature branch: git checkout -b feat/<feature-name>

Phase 2: Implement Feature

Skip if: User confirms feature code is already complete.

1. If frontend work: invoke frontend-design skill for UI quality
2. Implement the feature following existing project patterns
3. Write tests if test framework exists
4. Commit work: git add . && git commit -m "feat: <description>"

Phase 3: Local Test

Skip if: User confirms local testing passed.

1. Run project test suite if it exists
2. Start app locally and test with curl or equivalent
3. If frontend: use agent-browser to visually verify the UI
4. Fix any failures before proceeding

Phase 4: Docker + CI/CD

Skip if: Dockerfile and .github/workflows/deploy.yml already exist and are correct.

1. Create Dockerfile (multi-stage build, minimal final image)
2. Build and test locally: docker build -t <app> . && docker run -p 8080:8080 <app>
3. Create .github/workflows/deploy.yml:
   • Trigger on push to main
   • Build Docker image
   • Push to Azure Container Registry (ACR)
   • Deploy to Azure Container Apps
4. Commit CI/CD files and push branch
5. Open PR: gh pr create --title "feat: <description>" --body "<summary>"

Phase 5: CI/CD Monitor

Skip if: PR checks already passing.

1. Delegate to /aurora: "Monitor CI/CD pipeline for PR #N in <owner>/<repo>. Report status every 30s until complete."
2. If pipeline fails: read logs, fix issue, push fix, re-monitor
3. Once green: merge PR via gh pr merge --squash

Phase 6: Azure Deploy

Skip if: App already deployed and running at expected URL.

1. Ask user for: Azure resource group, ACR name, region (or use defaults)
2. Ensure Azure resources exist (create if needed):

   az group create --name <rg> --location <region>
   az acr create --name <acr> --resource-group <rg> --sku Basic
   az containerapp env create --name <env> --resource-group <rg>
   

3. Build and push image: az acr build --registry <acr> --image <app>:latest .
4. Deploy container app:

   az containerapp create --name <app> --resource-group <rg> \
     --environment <env> --image <acr>.azurecr.io/<app>:latest \
     --target-port 8080 --ingress external
   

5. Capture the live URL from output

Phase 7: Verify Deployment

Skip if: Live URL responds correctly.

1. Test live URL with curl — check HTTP 200
2. If frontend: use agent-browser to visually verify deployed UI
3. Delegate to /aurora: "Check logs for <app> in resource group <rg> — report any errors."
4. If verification fails: analyze logs, fix, push, wait for CI/CD, redeploy, re-verify (max 3 retries)
5. Present final summary: live URL, resource group, container app name

When to Stop and Ask

• Azure credentials not configured (az login needed)
• User hasn't specified which feature to build
• Deployment target ambiguous (Container Apps vs App Service vs ACI)
• Budget/cost concerns about Azure resources

Integration

• frontend-design — Invoke for any UI work in Phase 2
• superpowers:verification-before-completion — Run before declaring done
• /aurora — Delegate CI/CD monitoring (Phase 5) and Azure log checks (Phase 7)
• agent-browser — Visual testing in Phases 3 and 7