What problem does it solve? Working with Google Cloud Platform involves dozens of services, IAM subtleties, and operational pitfalls (startup probe failures, secret leaks in logs, Terraform state drift) that are easy to get wrong without deep operator experience. This Skill gives an AI agent a working operator's knowledge of GCP so it can run the right gcloud commands, follow least-privilege IAM patterns, and know when to escalate before making risky changes. ## Core Features & Use Cases - Task-based reference routing: A lookup table maps tasks (Cloud Run deploys, Cloud SQL migrations, IAM audits, Terraform multi-env, networking, storage) to focused reference files loaded on demand. - Day-to-day command patterns: Ready-to-use gcloud commands for authentication, log reading, Cloud Run inspection and rollback, IAM policy inspection, and Terraform workflows. - Debugging recipes: Step-by-step diagnostic flows for Cloud Run 5xx errors, infrastructure drift detection, cost investigation, and permission-denied IAM issues. - Project conventions auto-discovery: Walks up from the working directory to find a project-owned gcp-conventions.md file that overrides generic patterns. - Use Case: A Cloud Run service starts returning 503s after a deploy. The agent follows the debugging recipe: check revisions and traffic split, read recent ERROR logs, look for startup probe failures, and verify the runtime service account has secretmanager.secretAccessor on its secrets. ## Quick Start Ask the agent to diagnose why your Cloud Run service is returning 503 errors in your GCP project.
