claude

google-adk-skills

Skills and subagents for building production Google ADK agents.

1 on GitHubby OMIXECv1.1.1Apache-2.0updated 27 days ago
/

Plugin google-adk-skills from marketplace google-adk-skills in OMIXEC/Google-ADK-Skills. Commands read from the repo's manifest — 26 days ago.

This plugin runs code on your machine — it starts local MCP servers. The exact commands are listed below; read them before installing.

What it installs

Skills (35)

  • adk-git

    Writes commit messages and pull request descriptions for the adk-python repository: Conventional Commits types and scopes, subject lines that say why a change was made, and the linked-issue and testing-plan sections the PR template requires. Use when writing or rewording a commit message, squashing

  • adk-review

    Reviews the uncommitted changes in an adk-python working tree and reports correctness, design, public-API stability, test, sample and documentation gaps as a prioritized findings report, fixing them only when asked. Use when the user asks to review local changes, wants a self-review before opening a

  • adk-sample-creator

    Creates a new sample agent in the ADK Python repository — the sample directory, its `agent.py`, and its `README.md` — following the conventions the existing samples already use. Use when the user wants to add a sample or example demonstrating a feature or agent pattern (dynamic nodes, fan-out/fan-in

  • adk-setup

    Sets up a local ADK Python development environment in a git clone of the open-source adk-python repository: a uv virtual environment, all dependency extras, pre-commit hooks, and a first unit-test run. Runs only when explicitly requested, never on its own. Use when asked to set up, bootstrap, or rep

  • adk-style

    Python style and codebase conventions for ADK (Agent Development Kit): private-by-default file visibility, imports, type hints, Pydantic v2 models, formatting, docstrings, logging, async I/O, file and test layout, and unit test structure. Use when writing or editing ADK source or tests, deciding whe

  • adk-unit-design

    Writes an as-built architecture document for one ADK code unit — purpose, execution flow, data flow, cross-class dependencies, extension points, and the parts that must not change — to `docs/design/{topic}/{unit}/index.md`. It describes the code as implemented, not a proposed design, and its reader

  • adk-unit-guide

    Writes a hands-on developer guide for one ADK code unit — a minimal runnable example, how it works, a configuration-option table, advanced uses, limitations, and links to related samples — to `docs/guides/{topic}/{unit}/index.md`, then lists it in the index at `docs/guides/README.md`. Its reader is

  • adk-verify-snippets

    Checks that every Python code block in a Markdown file actually compiles and runs, by extracting each block to a temporary file, executing it in an isolated subprocess, and writing a pass/fail report with per-snippet coverage. Use when the user asks to verify, test, or validate the code samples in a

  • adk-a2a

    Implements the ADK Agent-to-Agent (A2A) protocol: agent cards, remote delegation over HTTP, streaming task exchange, and auth propagation across service boundaries. Use when agents in different processes or languages must call each other. Don't use for in-process sub-agent delegation (use adk-agent-

  • adk-agent-architecture

    Designs the internals of an ADK 2.x agent or cooperating set: agent modes, custom BaseAgent subclasses, tool boundaries, MCP and A2A integration, session state, context-window budget, memory, structured output, and model routing. Use when deciding what one agent is made of, which tools it may call,

  • adk-agent-builder

    Builds a single ADK Python agent from scratch: the LLM agent, its instruction, its tools, task and single-turn modes, and the first test. Use when creating one agent or adding a tool or a mode to an existing one. Don't use for how multiple nodes connect (use adk-graph-workflows), for deep tool, memo

  • adk-agent-orchestration

    Controls how an ADK agent runs over time: autonomous OODA loops with explicit planning and stop conditions, and LangGraph state machines for cyclic or branching control flow with ADK tools as nodes. Use when the agent must pursue a goal without per-step input, plan its own steps, drive an inspectabl

  • adk-agent-personas

    Instantiates a ready-made or bespoke persona agent: 30+ pre-built templates (tutor, coach, analyst, support rep, and others) plus a generator for domains with no matching template. Use when the request is about who the agent is: its role, expertise, interaction style, and scope. Don't use for how th

  • adk-agents

    Composes multiple ADK agents into a team: supervisor, router, sequential and parallel delegation, sub-agent handoff, and custom BaseAgent subclasses that control the run loop. Use when deciding which agent owns which job, how they delegate to each other, or when a prebuilt agent type cannot express

  • adk-architecture

    Explains how the ADK runtime fits together: the node and graph execution model, Context and Event flow, checkpoint and resume, tracing, and the rules governing the public API surface. Use when answering "how does X work" about ADK internals, tracing where an event or a piece of state comes from, or

  • adk-backend

    Builds the service layer around an ADK agent: HTTP server, request lifecycle, session endpoints, and persistence wiring. Use when exposing an agent behind an API or integrating it into an existing backend service. Don't use for the agent's tools or memory design (use adk-agent-architecture), for exe

  • adk-bidi-live

    Builds real-time ADK experiences on the Live API: bidirectional audio and video, WebSocket streaming, LiveRequestQueue, and Gemini 3.8 Live models. Use when building a voice agent, streaming chat, or multimodal live interaction, or when migrating a 2.5/3.1 Live model to gemini-3.8-live. Don't use fo

  • adk-configs

    Defines ADK agents declaratively: YAML agent definitions, environment variables, agent cards, and per-environment configuration. Use when the agent should be configured rather than coded, or when several environments share one definition. Don't use for programmatic agent construction (use adk-agent-

  • adk-debug

    Diagnoses a misbehaving ADK agent by inspecting sessions, events, tool calls, and the exact request that reached the model. Covers the adk run CLI, the adk web dev server, its session and trace endpoints, and a symptom catalogue for the common failure shapes. Use when an agent returns the wrong answ

  • adk-deploy-production

    Takes an ADK agent or workflow to production: CI/CD pipelines, container and cloud targets, identity and database integration, guardrails, observability, resilience, evaluation, and the pre-deploy checklist. Use when packaging an agent, wiring auth or a database, adding tracing and metrics, designin

  • adk-deployment

    Runs ADK's own deploy path: Cloud Run, GKE, Vertex AI Agent Engine, and the Docker and Kubernetes manifests behind them, using the adk deploy commands. Use when executing a deployment or writing a provider-specific manifest. Don't use for CI/CD, security, observability, evaluation, or the pre-deploy

  • adk-embeddings

    Generates and stores embeddings for ADK agents: multimodal embeddings with gemini-embedding-2, session and memory persistence in Cloud SQL, Postgres, Firestore, or SQLite, and vector search through pgvector or Firestore vector. Use when choosing a session or memory backend, wiring a database Session

  • adk-graph-workflows

    Designs and scaffolds ADK 2.x multi-agent workflows: graph, dynamic, collaborative, and template (sequential, parallel, loop) shapes, project layout, code generation, and extension of an existing workflow. Use when deciding how agents and deterministic steps connect, scaffolding a workflow project,

  • adk-litellm

    Routes ADK agents across 100+ model providers through LiteLLM: OpenAI, Anthropic, Bedrock, OpenRouter, Ollama, and vLLM, with provider switching, fallbacks, and cost trade-offs. Use when the agent must use a non-Google model or fail over across providers. Don't use for choosing among Gemini models (

  • adk-mcp

    Integrates the Model Context Protocol into ADK agents: McpToolset, stdio and streamable HTTP connections, tool discovery, filters, and MCP database toolboxes. Use when consuming an external MCP server or building MCP-based tools. Don't use for native function tools (use adk-tools), for agent-to-agen

  • adk-memory

    Manages conversation memory and state in ADK: session state, long-term memory, persistence backends, and cross-session recall. Use when implementing memory policy, persisting an agent's state, or deciding what survives a session. Don't use for database-backed session wiring and vector stores (use ad

  • adk-migration

    Migrates ADK code from 1.x to the 2.x line through 2.12.0: breaking API changes (Workflow versus GraphAgent, McpToolset, BuiltInCodeExecutor, ctx.resume_inputs, agent modes), renamed imports, deploy CLI changes, model and embedding updates, and session incompatibility. Use when upgrading an existing

  • adk-model-routing

    Selects and routes models for ADK agents: the current Gemini catalog (gemini-4-argon, gemini-3.8-flash, gemini-3.8-live, gemini-3.8-live-extended-thinking, gemini-3.1-pro-preview, gemini-embedding-2), thinking_level effort tiers, retired sampling parameters, redirecting model IDs, and primary-to-fal

  • adk-performance-scaling

    Makes an ADK agent faster and cheaper under load: latency decomposition, concurrent invocations, tool-thread and LLM-call limits, 429 and 5xx retry policy, context caching, event compaction, token budget, and provider capacity. Use when an agent is slow, times out, hits quota or rate limits under co

  • adk-prompts

    Writes and tunes ADK agent instructions: system prompts, few-shot examples, output-shape guidance, and token-cost trade-offs. Use when the agent's behavior is right in principle but wrong in wording or format. Don't use for choosing a persona (use adk-agent-personas), for tool schemas (use adk-tools

  • adk-rag

    Adds retrieval-augmented generation to ADK agents with Pinecone (self-managed) or Vertex AI RAG (managed corpora): ingestion, chunking, retrieval, and grounding. Use when the agent must answer from a document corpus or a knowledge base. Don't use for session or agent memory (use adk-memory), for raw

  • adk-runtime

    Explains and extends ADK execution internals: the event loop, Runner model, callbacks, and the Reason-Act cycle. Use when implementing custom callbacks or understanding why an invocation flowed the way it did. Don't use for the service layer around the agent (use adk-backend), for workflow topology

  • adk-skills-authoring

    Authors ADK agent skills: the SKILL.md contract, folder layout, frontmatter limits, the loader, SkillRegistry, and SkillToolset with its discovery and lifecycle modes. Use when creating a skill, deciding what belongs in SKILL.md versus references, wiring skills into an agent, or debugging a skill th

  • adk-tools

    Defines the tools an ADK agent may call: function tools, Google built-in tools such as Search and Code Execution, OpenAPI integrations, and tool composition. Use when writing a tool, wiring an external API, or deciding between tool types. Don't use for MCP servers (use adk-mcp), for debugging a tool

  • adk-typescript-go-parity

    Provides reference templates and parity notes for building ADK agents in Go and TypeScript. Use when the target runtime is Go or TypeScript, when porting a Python workflow to either, or when checking whether a feature exists outside Python. Don't use for Python agent or workflow work (use adk-graph-

Slash commands (5)

  • /google-adk-skills:adk-build-agent

    Build or modify a Google ADK agent using the bundled ADK skills

  • /google-adk-skills:adk-graph-visualize

    Render an ADK 2.3 Workflow's node/edge/join structure as a diagram

  • /google-adk-skills:adk-install-skills

    Install Google ADK skills into supported agent tools

  • /google-adk-skills:adk-migrate-check

    Scan the repo for lingering ADK v1 patterns and stale 2.x API-name drift

  • /google-adk-skills:adk-validate-skills

    Validate ADK skills, commands, agents, package metadata, and installer behavior

Subagents (8)

  • adk-agent-builder

    Build, test, and iterate on ADK agents: agent creation, mode configuration (task/single-turn), graph workflows with function nodes, routing/conditions, LLM agent nodes, RAG, personas, domain experts, and orchestration. Use when creating or configuring an ADK agent.

  • adk-architect

    Deep ADK architecture knowledge: graph orchestration, resumption/checkpoint, execution flow, node contracts, Runner/NodeRunner/Workflow separation, LLM context orchestration, and observability design. Use when understanding ADK internals, event flow, state management, or designing core components.

  • adk-debugger

    Debug ADK agents: inspect sessions, test agent behavior, troubleshoot tool calls, diagnose event flow and LLM/model issues. Supports `adk web` (browser UI + REST) and `adk run` (CLI) debugging modes.

  • adk-git-ops

    Git operations for ADK projects: commit, push, pull, rebase, branch, PR, cherry-pick. Enforces Conventional Commits with scope guidance. Use for any git operation on ADK repositories.

  • adk-setup

    Set up a local ADK Python development environment: Python 3.10+, the uv package manager, pre-commit hooks, and project dependencies. Use when getting started, setting up a new environment, or preparing to contribute.

  • adk-style-checker

    Enforce ADK code style and conventions: Python idioms, imports, typing, Pydantic patterns, formatting, logging, documentation, file organization, and testing rules. Use when writing code, reviewing PRs, or fixing style nits.

  • adk-workflow-builder

    Generate complete, production-grade ADK workflow projects from a confirmed architecture: agent definitions, tools, CI/CD, evals, security, observability, IaC, and deployment manifests. Python (default), Go, or TypeScript.

  • adk-workflow-designer

    Analyze requirements for multi-agent systems and propose ADK workflow architecture: agent roles, graph topology, coordination patterns, deployment target. Produces a structured architecture proposal before any code is written.

MCP servers (1)

  • context7

    runs: npx -y @upstash/context7-mcp

View source on GitHub

Similar plugins

Matt Pocock's agent skills for real engineering (grilling, spec/ticket flows, TDD, code review, domain modelling and more).

39 skills

mattpocock/skills 323,108

claude-mem

claudecodex

Persistent memory for Claude Code: the next session starts where the last one left off.

29 skills · 1 MCP server · hooks

thedotmack/claude-mem 116,618

mem0

claudecodex

Cross-session memory and token savings for coding agents.

6 skills · 1 agent · 1 MCP server · hooks

mem0ai/mem0 68,893

context7

claudecodex

Up-to-date documentation lookup. Pull version-specific documentation and code examples directly from source repositories into your coding agent's context.

1 MCP server

upstash/context7 64,709

Automates browser interactions for web testing, form filling, screenshots, and data extraction

12 skills

vercel-labs/agent-browser 50,373

agent-teams

claudecodex

Orchestrate multi-agent teams for parallel code review, hypothesis-driven debugging, and coordinated feature development using Claude Code's Agent Teams

6 skills · 7 commands · 4 agents

wshobson/agents 42,461

Maintain google-adk-skills?

This page updates itself from your repo. Add the badge to your README, or get it Featured to pin it above the organic results with a followed link to your site.

claude & codex pluginslisted