VividKit Crystal clear AI coding
VI
Choose Your Setup Get Started with AgentKit Continue with ClaudeKit Migrate from ClaudeKit to AgentKit Clean cutover from ClaudeKit What Is AgentKit? How AgentKit Works AgentKit Kits, Skills, Agents & Hooks AgentKit Runtime Support AgentKit Desktop App AgentKit Configuration AgentKit CLI Quick Reference AgentKit Skills Cheatsheet AgentKit Workflows AgentKit Statusline Update AgentKit CLI & Kits AgentKit Helper Clean reinstall of AgentKit AgentKit Troubleshooting What is ClaudeKit? ClaudeKit CLI Guide Command Finder Commands Cheat Sheet Workflow Recipes CLI Quick Reference Claude Mechanics Permission Modes ClaudeKit Hooks Coexistence ClaudeKit x Codex CK → Other coding tools How ClaudeKit Works Inside ClaudeKit Finding Your Unknowns Session Recovery Fix from Logs UI/UX Guide Accounts Switcher IDE Config Remote Control Codex App System Diagramming UI Review Gate Promotions Support VividKit

Quick Ref / Command

Engineer Kit

/ak:google-adk-python

Google ADK Python Builder

Guide code-first Google ADK Python work: agents, workflow agents, A2A, MCP tools, state, memory, callbacks, plugins, evaluation, and deployment.

01

Choose target

02

Set structure

03

Pick pattern

04

Wire tools

Rule 01

ADK agents are code-first Python artifacts that should be versioned and tested.

Rule 02

Follow the root_agent/app convention so ADK CLIs can discover the agent.

Rule 03

Use workflow agents when control flow is predictable; use LLM agents for dynamic routing.

Execution Map

Core lanes from input to output, aligned with the command's real execution path.

01

Start

Input, scope, route

  1. 1 Choose target Identify whether the user needs a single agent, multi-agent system, workflow pipeline, A2A remote agent, MCP integration, evaluation, or deployment.
  2. 2 Set structure Create the required my_agent package shape with __init__.py importing agent and agent.py exposing root_agent or app.
02

Work

Agent / skill execution

  1. 3 Pick pattern Use Agent/LlmAgent for dynamic routing, SequentialAgent for fixed order, ParallelAgent for concurrency, LoopAgent for iteration, or RemoteA2aAgent for A2A.
  2. 4 Wire tools Attach Python tools, MCPToolset, sub_agents, LongRunningFunctionTool, state, artifacts, callbacks, or plugins as required by the feature.
03

Verify

Gate, review, validation

  1. 5 Use App when needed Switch from root_agent-only to App when plugins, event compaction, file artifacts, or custom lifecycle management are needed.
  2. 6 Exercise locally Use adk web for the dev UI, adk run for CLI testing, or adk api_server for FastAPI serving.
04

Close

Report, handoff, artifact

  1. 7 Evaluate Run adk eval with an evalset JSON before treating the agent as production-ready.
  2. 8 Deploy consciously For production, route to Cloud Run, Vertex AI Agent Engine, or GKE guidance and keep model/provider assumptions explicit.

Syntax, arguments, subcommands

How to invoke the skill: syntax, positional arguments, shared options, then each subcommand with its own syntax and outcome.

Syntax

/ak:google-adk-python [agent or feature]

Arguments

[agent or feature] Required

ADK agent or feature

Natural-language outcome and operating boundaries for a Google ADK Python agent: target package, agent responsibility, tools or MCP access, state or memory needs, evaluation expectations, provider limits, and whether deployment is in scope. This is not an ADK CLI subcommand or mode flag.

/ak:google-adk-python "Add a Google ADK Python support agent to this existing package. Use the installed SDK, expose read-only account tools through a filtered MCPToolset, keep approval for any write, add a small eval set and unit tests, and do not call a paid model or deploy."

Sample Prompt

Concrete invocations for each flag, subcommand, mode, or route available in the live workflow.

Tool-using support agent Recommended 
/ak:google-adk-python build a support triage agent with MCP tools, session state, and artifact output
Use when:
You are building a Google ADK agent that needs MCP tools, state, or artifact handling.
Expected:
Defines the ADK package shape, chooses root_agent or App, wires MCPToolset and ToolContext state, and names the adk web/run/eval checks to exercise the agent.
Workflow pipeline 
/ak:google-adk-python design a SequentialAgent pipeline with a ParallelAgent research step and LoopAgent refinement
Use when:
You need workflow agents for predictable sequential, parallel, or iterative processing.
Expected:
Maps the pipeline to SequentialAgent, ParallelAgent, and LoopAgent roles, explains when not to use dynamic LlmAgent routing, and keeps the agent.py export discoverable by ADK.
A2A coordinator 
/ak:google-adk-python design a RemoteA2aAgent coordinator for two remote domain agents
Use when:
You are connecting remote agents through the A2A protocol instead of local sub-agents.
Expected:
Explains the RemoteA2aAgent coordinator role, how it differs from local sub_agents composition, and what local CLI or deployment checks should prove the connection path.
Eval and Vertex deployment 
/ak:google-adk-python prepare my ADK agent for adk eval and Vertex AI Agent Engine deployment
Use when:
You need to harden an ADK agent with evaluation and production deployment guidance.
Expected:
Lays out the evalset JSON and adk eval path, confirms whether App is needed for plugins or lifecycle handling, and routes production deployment to Cloud Run, Vertex AI Agent Engine, or GKE.

Handled Scope

  • Agent and App structure
  • MCP tool integration
  • A2A multi-agent systems
  • State, memory, artifacts, callbacks, plugins
  • ADK eval and Vertex/Cloud Run deployment

Artifacts Produced

ADK implementation guidance

Pattern

Recommended agent type, package structure, key APIs, CLI checks, eval/deploy path.

The result should leave the user with an ADK-discoverable Python agent or a concrete implementation plan.

Next