Cloud Run Deployment Guide
Deploy and manage Cloud Run services, jobs, and worker pools with gcloud.
The documentation identifies IAM roles, service identities, Secret Manager, ingress controls, least-privilege guidance, and container security practices, and it flags public-access risks. However, commands use broad roles, --quiet, and an --allow-unauthenticated example without per-action confirmation, least-privilege defaults, explicit data-flow disclosure, or a concrete rollback/recovery procedure, so points are deducted.
The main service, job, and worker-pool paths have reasonably complete commands and some failure troubleshooting. Static evidence includes no tests or execution results; several examples omit important project or region context, the CLI reference contains a visibly malformed line-broken function deployment command, and abnormal-input and failure-feedback coverage is limited, so points are deducted.
The intended scenarios, three Cloud Run resource types, and broad boundaries are clear, with CLI, client-library, Terraform, and MCP coverage. Trigger conditions, input/output contracts, and environment prerequisites are not systematically defined; Chinese-language support is absent, and reachability from mainland-China networks is not addressed despite dependence on Google Cloud services, so points are deducted.
The skill uses a layered main document plus references and includes prerequisites, examples, security, networking, cost, and troubleshooting material. The repository supplies Apache-2.0 licensing, installation, contribution, and issue-reporting paths and states that it is under active development. The skill lacks its own version, changelog, explicit maintenance owner, compatibility policy, and FAQ, so points are deducted.
The material can guide common deployment and management tasks for Cloud Run services, jobs, and worker pools, and the examples provide usable starting points. Static review cannot verify commands or outputs; examples depend on substantial IAM, API, region, image, and project configuration, and some may require manual correction, so the static ceiling and limited direct-usability evidence constrain the score.
Most key claims link to first-party Google Cloud, Terraform, SDK, or MCP documentation, making the source auditable. There is no committed test suite, CI execution evidence, or independent reproduction evidence in the supplied material, and linked content was not verified during this static review, so points are deducted.
- Before execution, confirm the project, region, image provenance, IAM scope, and cost impact; do not default to public unauthenticated access or overly broad service-account permissions.
- Failure guidance points to logs but does not define a clear revision rollback, traffic-recovery, or resource-cleanup procedure.
- Some CLI examples require manual verification, and mainland-China network reachability for the core services is not addressed.
What does this skill do, and when should you use it?
This skill helps users run code, functions, or containers on Google Cloud Run through its three resource types: services, jobs, and worker pools. It covers API enablement, IAM setup, deployment from container images or source code, and job execution. It also gives troubleshooting guidance for permission errors, boot crashes, health-check failures, and native dependencies. Adoption requires a Google Cloud project, appropriate permissions, and an available gcloud environment.
Provides instructions for enabling the Cloud Run Admin and Cloud Build APIs and assigning Cloud Run Admin, Cloud Run Source Developer, Service Account User, Logs Viewer, and Cloud Run Builder roles; deploys services and worker pools from Artifact Registry, Docker Hub, or configured remote registries; builds and deploys services or worker pools from source; creates, configures, executes, and waits for Cloud Run jobs; reads Cloud Logging output when deployments fail; and points to references for client libraries, MCP, Terraform, IAM, security, networking, and cost optimization.
- A developer needs to deploy a stateless HTTP application with autoscaling instances and a stable endpoint.
- A team needs to run parallel tasks manually or on a schedule until completion.
- An operations team needs an always-on pull-based workload such as a Kafka, Pub/Sub, or RabbitMQ consumer.
- A team wants to deploy from source using Cloud Build and buildpacks without installing Docker locally.
- An engineer needs to diagnose IAM errors, boot crashes, health-check failures, or native dependency problems after deployment.
What are this skill's strengths and limitations?
- Covers services, jobs, and worker pools in one Cloud Run skill.
- Includes copyable gcloud commands for image deployment, source deployment, job execution, and troubleshooting.
- Clearly identifies deployment IAM roles and links to buildpacks, Cloud Build, Terraform, and networking references.
- Requires a Google Cloud project and several IAM permissions; it does not bypass cloud access controls.
- The source provides no cost estimates, test suite, or concrete client compatibility validation.
- It depends on Google Cloud services such as Cloud Run, Cloud Build, and Artifact Registry; source deployment without build is marked Preview.
How do you install this skill?
In an Agent Skills-compatible environment, run npx skills add google/skills, then select skills/cloud/cloud-run-basics from the repository. The supplied source does not document a client-specific installation directory or a separate installation command for this skill.
How do you use this skill?
Ask an agent: “Using Cloud Run Basics, help me deploy a Cloud Run service from the current directory and verify the required APIs, IAM roles, and port configuration.” For deployment, first run gcloud services enable run.googleapis.com cloudbuild.googleapis.com --quiet, then use gcloud run deploy, gcloud run jobs create, or gcloud run worker-pools deploy as appropriate. Deployed code must listen on 0.0.0.0 and use the injected $PORT variable.
How does this skill compare with similar options?
The skill distinguishes three Cloud Run resource types and two deployment paths: services handle requests, jobs run to completion, and worker pools support always-on pull-based workloads; deployments can use a container image or source code built through Cloud Build/buildpacks.