Google Cloud Agent Skills

google/skills is Google's Agent Skills catalog for Google products and technologies, including Google Cloud and Gemini Enterprise Agent Platform. It lets coding agents install focused guidance for Gemini API, Agent Platform Skill Registry, Cloud Run, BigQuery, Firebase, GKE, Cloud SQL, AlloyDB, Workload Manager, gcloud, authentication, onboarding, network observability, and Google Cloud Well-Architected Framework reviews.

This entry is separate from Google Agents CLI Skills. google/agents-cli gives agents an end-to-end ADK build/eval/deploy lifecycle, while google/skills provides product-specific Google Cloud and Agent Platform capability packs that can be installed alongside any other coding-agent workflow.

Knowledge Freshness

The README says the repository is under active development. Its current catalog includes Agent Platform APIs, Gemini API guidance, Google Cloud basics, auth/onboarding recipes, network observability, and Well-Architected Framework skills. Google Cloud APIs, product names, SDK recommendations, launch stages, IAM roles, pricing, quotas, regions, model names, and CLI flags change often.

Use these skills for source-backed Google workflow guidance, then verify exact commands, flags, permissions, data movement, and production constraints against the target Google Cloud project and current official docs before running changes.

Retrieval Sources

This listing is grounded in:

• The google/skills repository README.
• The public skills.sh listing for google/skills.
• Representative SKILL.md files for Agent Platform Skill Registry, Gemini API, gcloud, Cloud Run, and BigQuery.
• The Apache-2.0 license text.
• Current GitHub repository metadata.

Core Workflow

Install the Google skills catalog:

npx skills add google/skills

Then choose the focused skill or skills for the task. For example:

• Gemini API on Agent Platform for multimodal generation, function calling, structured output, embeddings, context caching, Live API, and batch prediction.
• Skill Registry for discovering, uploading, updating, and managing Agent Platform skills.
• gcloud for safe Google Cloud CLI usage with command validation and data reduction rules.
• Cloud Run, BigQuery, Firebase, GKE, Cloud SQL, AlloyDB, Workload Manager, auth recipes, onboarding, network observability, or Well-Architected review.

Capability Scope

The README currently lists skills across these areas:

The catalog also links to adjacent official Flutter and Dart skills repositories for teams working across Google application stacks.

Production Rules

Use these skills with strict cloud-change boundaries:

• Start with the narrowest product skill that matches the task.
• Validate Google Cloud authentication, target project, location, enabled APIs, IAM roles, billing, and quotas before running commands.
• Use current command help and official docs for CLI syntax instead of relying on memory.
• Require explicit approval before create, update, delete, deploy, IAM, network, billing, data export, registry, or production-impacting operations.
• Prefer read-only discovery with reduced output before making changes.
• Treat Agent Platform, Gemini API, BigQuery, Cloud Run, Firebase, GKE, Cloud SQL, AlloyDB, and Terraform actions as operational changes, not harmless examples.
• Record what project, region, resource, command, and validation result was used without leaking credentials or private data.

Use Cases

• Ask Codex to choose the right Gemini API setup for Agent Platform and avoid deprecated SDKs.
• Use Claude Code to inspect a Cloud Run deployment plan before running gcloud commands.
• Have an agent write or review BigQuery SQL, data ingestion, MCP usage, IAM, and AI/ML workflow guidance from source-backed skill references.
• Use Gemini CLI to prepare Google Cloud authentication and onboarding steps for a new project.
• Ask an agent to run a Well-Architected review for security, reliability, cost, performance, operations, or sustainability.

Source Review

• The README describes the repository as Agent Skills for Google products and technologies, including Google Cloud.
• The README documents installation with npx skills add google/skills.
• The available skills list includes Gemini API on Agent Platform, Gemini Interactions API, Managed Agents API, Skill Registry API, AlloyDB, BigQuery, Cloud Run, Cloud SQL, Firebase, GKE, Workload Manager, auth/onboarding recipes, network observability, and Well-Architected pillars.
• The Agent Platform Skill Registry skill covers skill discovery, lifecycle management, operation monitoring, and local skill generation.
• The Gemini API skill requires the unified Google Gen AI SDK and warns against legacy SDKs.
• The gcloud skill requires leaf-command validation, data reduction, explicit project/location scoping, non-interactive execution, and destructive-action authorization.
• The Cloud Run and BigQuery skills provide concrete prerequisites, roles, commands, and product-specific references.
• The license file identifies the repository as Apache-2.0.

Duplicate Review

Checked current content/skills/, content/tools/, content/mcp/, open pull requests, and repository-wide content for google/skills, Google Cloud Agent Skills, Gemini API skills, Agent Platform Skill Registry, Cloud Run skills, BigQuery skills, Firebase skills, GKE skills, and gcloud skills. Existing Google entries cover separate tools and MCP servers, including Google ADK, Firebase MCP, and Google Agents CLI; no dedicated google/skills capability pack, exact source URL duplicate, target file, or open duplicate PR was found.

Disclosure

Editorial listing. No paid placement or affiliate link is used. The upstream repository is Apache-2.0 licensed and states that it is under active development. Google Cloud, Gemini Enterprise Agent Platform, Gemini APIs, Firebase, BigQuery, Cloud Run, GKE, Cloud SQL, AlloyDB, and related services may have separate billing, terms, IAM, privacy, residency, retention, and launch stage constraints.