adk-agent-builder

A building guide for ADK (Agent Development Kit) Python agents and workflows. It covers language-model agents, tools, graph routing, parallel work, delegation, pauses for human input, and pytest tests.

In plain words
What is it for?
Use it to create an agent, add tools, build routed or parallel workflows, delegate tasks between agents, handle human review, and test the result.
Why use it?
It provides the relevant implementation rules for creating agents without guessing at APIs or workflow behavior.

Skill for Claude CodeCodex

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/google/adk-python/adk-agent-builder
Any agent
npx skills add google/adk-python --skill adk-agent-builder
Clone the repo
git clone --depth 1 https://github.com/google/adk-python

Made for: Claude Code, Codex.

Per session 177 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 884 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00177 $0.00884
Opus 5 $0.00088 $0.00442
Sonnet 5 $0.00035 $0.00177
Haiku 4.5 $0.00018 $0.00088

Measured yesterday against content hash 5690ecc21fa4, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

adk-agent-builder scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

.agents/skills/adk-agent-builder/SKILL.md · 72 lines

How it starts

The opening of the file, as written. The whole thing — 72 lines — stays where its author put it; the contents beside it link to each section on GitHub.

ADK Agent Builder

Read only the reference that matches the task. Loading the whole tree costs context and buries the part that matters.

Every API below was checked against google-adk 2.6.2. If a symbol is missing at runtime, read the source under src/google/adk/ rather than guessing a neighbouring name.

Start here

Task Reference
First agent, environment, adk CLI getting-started.md
Which import path is the canonical one import-paths.md
The rules that cause most runtime failures best-practices.md

Building blocks

  • tool-catalog.md — function tools, MCP, OpenAPI, Google API toolsets, built-in tools, custom BaseTool and BaseToolset.
  • function-nodes.md — plain functions as nodes: parameter resolution, generators, node_input typing rules.
  • llm-agent-nodes.md — an LlmAgent used as a workflow node: output types, instruction templates, output_schema, auto-wrapping behavior.
  • task-mode.mdmode='task' and mode='single_turn' delegation with schema-validated input and output.

Graph orchestration

  • routing-and-conditions.md — routed edges, dict routing maps, default routes, self-loops, revision loops.
  • parallel-and-fanout.md — fan-out edges, JoinNode fan-in, parallel_worker=True list processing.
  • dynamic-nodes.md — scheduling nodes at runtime with ctx.run_node() and imperative workflow construction.
  • human-in-the-loop.mdRequestInput, resume behavior, resumable vs replayed sessions.
  • advanced-patterns.md — nested workflows, retries, custom BaseNode subclasses, graph validation rules.
  • multi-agent.md — chat-transfer hierarchies, and the deprecated SequentialAgent / LoopAgent / ParallelAgent shells that Workflow replaces.

Read the full file on GitHub · 72 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. yesterday First seen · 72 lines · 177 tokens per session scan A 5690ecc21fa4

Subscribe to this mod's changes

adk-agent-builder is a skill published in the GitHub repository google/adk-python (21,331 stars, last pushed 2d ago), licensed Apache-2.0. It adds 177 tokens to every session and 884 once invoked, about $0.0009 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.