Google Cloud Platform

**Judgement:** High-utility, broad-coverage GCP operations skill built on official CLIs. Very powerful for deployments, infrastructure management, and troubleshooting across multiple GCP services, but carries meaningful risk of downtime or cost impact if misused. **What it does well** - Covers a wide range of GCP services: Compute Engine, Cloud Run, Firebase Hosting, Cloud Storage, Secret Manager, Cloud SQL, Artifact Registry, Logging/Monitoring, and Billing. - Uses standard, well-documented commands from `gcloud`, `gsutil`, and `firebase-tools`, which are reliable and familiar to GCP practitioners. - Includes setup/authentication, logging/monitoring, and troubleshooting patterns, so it’s useful for both deployment and ongoing operations. **Key risks / cautions** - Many commands modify live infrastructure (start/stop/reset VMs, deploy/rollback services, modify traffic split, change IAM, create budgets, manage secrets, delete images, destroy secret versions). Misuse can cause downtime, misconfiguration, or data loss. - Cost risk: leaving resources running, misdirected deployments, or incorrect projects/regions can incur unexpected charges. - No apparent built-in guardrails or confirmation mechanisms beyond the CLI’s own prompts; safety depends heavily on correct parameters and responsible use. **Recommended scenarios** - Teams that already use GCP heavily and want LLM-assisted automation for deployments, operations, and debugging, especially SRE/DevOps workflows. - Power users who understand GCP resource scoping (project/region/zone) and want scripted/agent-driven management via CLIs. - Best used in: - Non-production or lower-risk environments while prompts and workflows are being refined. - Production environments only under strict governance, clear scoping (e.g., per-project), and human review of high-impact commands (deletes, rollbacks, secret changes, traffic changes).