Skip to content
agentusehub
Sign in+ Submit
Software & DevOpsAny

Scaffold a new service end to end with agents

Cursor agents create Terraform, k8s manifests and CI, validate them and open a draft PR; one cleanup run wrongly deleted Terraform files.

Read the experience
Time saved
Not reported
Setup time
Not reported
Running cost
Not reported
Cost range
Not reported

Use case overview

Cursor agents create Terraform, k8s manifests and CI, validate them and open a draft PR; one cleanup run wrongly deleted Terraform files.

Read the original report ↗

Reported experience

  1. What they did

    A task file in the repo, run with a service name and team name, drives the agent. It first checks GitHub, the cluster and Terraform state for conflicts. It then writes the Terraform service file, the Kubernetes Deployment, Service and HPA, and a CI workflow with a Trivy scan. It runs terraform validate and kubectl dry-run, then opens a draft PR with a checklist. A senior engineer reviews before merge.

  2. How it went

    The article describes a new engineer producing convention-compliant files this way, with no timing stated for this run. Separately, a vague prompt to clean up unused resources deleted 14 files, three of them Terraform files still managing state. A checkpoint rollback fixed it.

  3. Worth knowing

    Replace vague prompts like cleanup with precise report-only tasks. Reviewers who skimmed let a cpu=10m limit through, so keep PRs small and use checkpoints.

Example prompts

Try with Cursor
Scaffold a new service called [service name] including Terraform, Kubernetes manifests, and CI config, validating each file as you go. Open a draft pull request with the changes, but show me exactly what will be created or deleted before running any cleanup or deletion commands.

Related use cases

Software & DevOps

Port a Chrome extension to Safari

The author had ChatGPT take the Session Buddy extension source from Chrome, port it to Safari and install it, taking about 20 minutes.

ChatGPT