home / skills / cnemri / google-genai-skills / google-adk-python

google-adk-python skill

safe

This skill provides expert guidance and Python examples for building Google ADK agents, tools, streaming, and deployment workflows.

npx playbooks add skill cnemri/google-genai-skills --skill google-adk-python

Review the files below or copy the command above to add this skill to your agents.

Files (13)

SKILL.md

3.0 KB

---
name: google-adk-python
description: Expert guidance on the Google Agent Development Kit (ADK) for Python. Use this skill when the user asks about building agents, using tools, streaming, callbacks, tutorials, deployment, or advanced architecture with the Google ADK in Python.
---

# Google ADK (Python) Skill

This skill provides comprehensive documentation and Python examples for the Google Agent Development Kit (ADK). It maps documentation topics to their corresponding Python code snippets.

## How to Use

Identify the user's specific interest or task and refer to the relevant reference file below. Each reference file contains links to the official documentation (Markdown) and the corresponding Python examples (raw code).

## Topics

### 1. Getting Started
For installation, quickstarts, and basic agent setup.
- **Reference**: [references/get-started.md](references/get-started.md)

### 2. Agents & Models
For creating different types of agents (LLM, Workflow, Loop, Parallel, Sequential) and configuring specific models (Gemini, Anthropic, etc.).
- **Agents Reference**: [references/agents.md](references/agents.md)
- **Models Reference**: [references/models.md](references/models.md)

### 3. Tools (Basic & Advanced)
For integrating tools like Google Search, Code Execution, BigQuery, third-party services (GitHub, Jira, etc.), MCP, and Grounding.
- **Basic Tools Reference**: [references/tools.md](references/tools.md)
- **Advanced Tools Reference**: [references/advanced-tools.md](references/advanced-tools.md)

### 4. Streaming
For building real-time, low-latency streaming agents (audio/video).
- **Reference**: [references/streaming.md](references/streaming.md)

### 5. Callbacks
For hooking into agent lifecycle events (before/after agent, model, tool execution).
- **Reference**: [references/callbacks.md](references/callbacks.md)

### 6. Runtime & Architecture
For deep dives into the Runtime, Sessions, Memory, Context, Events, Artifacts, and Plugins.
- **Reference**: [references/runtime-arch.md](references/runtime-arch.md)

### 7. Deployment & Operations
For deploying agents (Cloud Run, GKE) and observability (Logging, Tracing, Evaluation).
- **Reference**: [references/deploy-ops.md](references/deploy-ops.md)

### 8. Tutorials & Samples
For end-to-end tutorials and complete agent samples (e.g., YouTube Shorts Assistant, Weather Agent).
- **Reference**: [references/tutorials.md](references/tutorials.md)

### 9. API Reference
For REST API details.
- **Reference**: [references/api.md](references/api.md)

### 10. General Information
For project info, community, release notes, and limitations.
- **Reference**: [references/general.md](references/general.md)

## Instructions

- When a user asks about a specific topic, load the corresponding reference file to get the URLs for the documentation and code.
- You can read the content of the linked files using `web_fetch` or `run_shell_command` with `curl` if you need to provide the actual content to the user.
- Always prefer providing the Python example code when explaining a concept.

Overview

This skill provides expert guidance and ready-to-run Python examples for building agents with the Google Agent Development Kit (ADK). It maps common ADK topics—agents, tools, streaming, callbacks, runtime, and deployment—to concise Python samples and official docs. Use it to accelerate development, avoid common pitfalls, and learn best practices for production-ready ADK agents.

How this skill works

I organize ADK topics into focused reference sections and pair each concept with Python example code you can run or adapt. When you ask about a specific task (e.g., tools, streaming, callbacks), I point to the relevant reference and provide the matching Python snippets and links to official docs. I also explain how to fetch and run examples locally when deeper inspection is needed.

When to use it

Building a new ADK agent (LLM, Workflow, Loop, Parallel, Sequential) in Python
Integrating tools: Google Search, BigQuery, code execution, or third-party services
Implementing real-time streaming for audio/video agents
Hooking callbacks to observe or modify agent lifecycle events
Designing runtime, session, memory, and plugin architectures
Deploying agents to Cloud Run, GKE, or adding observability and tracing

Best practices

Prefer the provided Python examples as starting templates and adapt configuration for your model and environment
Use callbacks for observability: log before/after model and tool calls to aid debugging and evaluation
Isolate sensitive credentials and use environment-based secrets for deployment
Start with small, testable tools and add grounding or MCP only after stable behavior is observed
Design sessions and memory with explicit size and eviction policies to keep context costs predictable

Example use cases

Quickstart: install ADK and spin up a basic LLM agent using the Python example
Tool integration: connect BigQuery or GitHub and run code-execution samples to validate tool outputs
Streaming agent: build a low-latency audio/video assistant using streaming examples
Callbacks: add lifecycle hooks to capture inputs, outputs, and metrics for evaluation
Deployment: package a tested agent and deploy to Cloud Run with logging and tracing enabled

FAQ

Do you provide runnable Python examples?

Yes. Each topic is paired with concise, runnable Python snippets that you can download and execute locally or in CI.

How do I find examples for a specific ADK feature?

Tell me the feature (agents, tools, streaming, callbacks, runtime, deployment) and I will surface the matching Python reference and the official documentation links.