google-agents-cli-adk-code

Provides ADK Python API patterns and code references for building agents, tools, and workflows.

Updated Jun 27, 2026
One-click install
npx skills add https://github.com/God-Gamer-Manyu/Google_Agentic_Engineering --skill google-agents-cli-adk-code-god-gamer-manyu
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: google-agents-cli-adk-code
Source: https://github.com/God-Gamer-Manyu/Google_Agentic_Engineering/tree/main/.agents/skills/google-agents-cli-adk-code
Command: npx skills add https://github.com/God-Gamer-Manyu/Google_Agentic_Engineering --skill google-agents-cli-adk-code-god-gamer-manyu

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Writing agents with Google's Agent Development Kit requires knowing the correct Python API patterns for agents, tools, callbacks, state management, and graph workflows, and this Skill supplies verified cheatsheets so you avoid incorrect imports and outdated signatures. ## Core Features & Use Cases - ADK Python Cheatsheet: Covers Agent definitions, FunctionTools, callbacks, plugins, state prefixes, artifacts, memory, MCP tools, and multi-agent orchestration with SequentialAgent, ParallelAgent, and LoopAgent. - Workflow API Reference: Documents the ADK 2.0 graph-based Workflow API including nodes, edges, fan-out/fan-in, human-in-the-loop interrupts, retry configuration, and parallel workers. - Use Case: When asked to add a tool with confirmation to an existing ADK agent, consult the reference to write a FunctionTool with require_confirmation and correct ToolContext state access. ## Quick Start Ask the assistant to write an ADK agent with a custom tool and callbacks using the patterns from this skill's references.

Frequently Asked Questions about google-agents-cli-adk-code

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I define an agent with tools in Google ADK Python?

Create an Agent with name, model, instruction, and a tools list containing Python functions with type hints and docstrings. The function's docstring is sent to the LLM, and it must return a JSON-serializable dict.

How do I build a multi-agent system with ADK?

Use SequentialAgent, ParallelAgent, or LoopAgent for deterministic orchestration, or give a coordinator agent sub_agents for LLM-driven delegation. You can also wrap an agent with AgentTool to invoke it as a tool while the parent stays in control.

Does ADK support human-in-the-loop approval?

Yes, enable ResumabilityConfig on the App, then use the request_input tool, FunctionTool with require_confirmation, or yield RequestInput from a workflow node. User replies are read from ctx.resume_inputs keyed by interrupt_id.

Can I use MCP servers as tools in ADK agents?

Yes, use McpToolset with StdioConnectionParams for local servers or SseConnectionParams for remote production servers. Paths must be absolute, and you can restrict exposed tools with the tool_filter parameter.

What are the limitations of the ADK Workflow API?

Workflows require google-adk 2.0 and Python 3.11 or higher, and do not support live streaming via Runner.run_live. Graphs also reject unconditional cycles and require all nodes to be reachable from START.