Haiyuan CaoandCopybara-Service
4260ef0c7c
This CL adds four enhancements to the BigQuery Agent Analytics plugin and fixes a span hierarchy corruption bug. - **Schema Auto-Upgrade:** Additive-only schema migration that automatically adds missing columns to existing BQ tables on startup. A `adk_schema_version` label on the table (starting at `"1"`, bumped with each schema change) makes the check idempotent — the diff runs at most once per version. Enabled by default (`auto_schema_upgrade=True`) because upgrades are additive-only and fail-safe. Pre-versioning tables (no label) are treated as outdated, diffed, and stamped. No previous schema versions need to be stored; the logic diffs actual columns against the current canonical schema. - **Tool Provenance:** Adds `tool_origin` to TOOL_* event content, distinguishing six origin types — `LOCAL` (FunctionTool), `MCP` (McpTool), `A2A` (AgentTool wrapping RemoteA2aAgent), `SUB_AGENT` (AgentTool), `TRANSFER_AGENT` (TransferToAgentTool), and `UNKNOWN` (fallback) — via `isinstance()` checks with lazy imports to avoid circular dependencies. - **HITL Tracing:** Emits dedicated HITL event types (`HITL_CONFIRMATION_REQUEST`, `HITL_CREDENTIAL_REQUEST`, `HITL_INPUT_REQUEST` + `_COMPLETED` variants) for human-in-the-loop interactions. Detection lives in `on_event_callback` (for synthetic `adk_request_*` FunctionCall events emitted by the framework) and `on_user_message_callback` (for `adk_request_*` FunctionResponse completions sent by the user), not in tool callbacks — because `adk_request_*` names are synthetic function calls that bypass `before_tool_callback`/`after_tool_callback` entirely. - **Span Hierarchy Fix (#4561):** Removes `context.attach()`/`context.detach()` calls from `TraceManager.push_span()`, `attach_current_span()`, and `pop_span()`. The plugin was injecting its spans into the shared OTel context, which corrupted the framework's span hierarchy when an external exporter (e.g. `opentelemetry-instrumentation-vertexai`) was active — causing `call_llm` to be re-parented under `llm_request` and parent spans to show shorter durations than children. The plugin now tracks span_id/parent_span_id via its internal contextvar stack without mutating ambient OTel context. Co-authored-by: Haiyuan Cao <haiyuan@google.com> PiperOrigin-RevId: 873114688
Agent Development Kit (ADK)
An open-source, code-first Python framework for building, evaluating, and deploying sophisticated AI agents with flexibility and control.
Important Links: Docs, Samples, Java ADK, Go ADK & ADK Web.
</html>Agent Development Kit (ADK) is a flexible and modular framework that applies software development principles to AI agent creation. It is designed to simplify building, deploying, and orchestrating agent workflows, from simple tasks to complex systems. While optimized for Gemini, ADK is model-agnostic, deployment-agnostic, and compatible with other frameworks.
🔥 What's new
-
Custom Service Registration: Add a service registry to provide a generic way to register custom service implementations to be used in FastAPI server. See short instruction. (391628f)
-
Rewind: Add the ability to rewind a session to before a previous invocation (9dce06f).
-
New CodeExecutor: Introduces a new AgentEngineSandboxCodeExecutor class that supports executing agent-generated code using the Vertex AI Code Execution Sandbox API (ee39a89)
✨ Key Features
-
Rich Tool Ecosystem: Utilize pre-built tools, custom functions, OpenAPI specs, MCP tools or integrate existing tools to give agents diverse capabilities, all for tight integration with the Google ecosystem.
-
Code-First Development: Define agent logic, tools, and orchestration directly in Python for ultimate flexibility, testability, and versioning.
-
Agent Config: Build agents without code. Check out the Agent Config feature.
-
Tool Confirmation: A tool confirmation flow(HITL) that can guard tool execution with explicit confirmation and custom input.
-
Modular Multi-Agent Systems: Design scalable applications by composing multiple specialized agents into flexible hierarchies.
-
Deploy Anywhere: Easily containerize and deploy agents on Cloud Run or scale seamlessly with Vertex AI Agent Engine.
🚀 Installation
Stable Release (Recommended)
You can install the latest stable version of ADK using pip:
pip install google-adk
The release cadence is roughly bi-weekly.
This version is recommended for most users as it represents the most recent official release.
Development Version
Bug fixes and new features are merged into the main branch on GitHub first. If you need access to changes that haven't been included in an official PyPI release yet, you can install directly from the main branch:
pip install git+https://github.com/google/adk-python.git@main
Note: The development version is built directly from the latest code commits. While it includes the newest fixes and features, it may also contain experimental changes or bugs not present in the stable release. Use it primarily for testing upcoming changes or accessing critical fixes before they are officially released.
🤖 Agent2Agent (A2A) Protocol and ADK Integration
For remote agent-to-agent communication, ADK integrates with the A2A protocol. See this example for how they can work together.
📚 Documentation
Explore the full documentation for detailed guides on building, evaluating, and deploying agents:
🏁 Feature Highlight
Define a single agent:
from google.adk.agents import Agent
from google.adk.tools import google_search
root_agent = Agent(
name="search_assistant",
model="gemini-2.5-flash", # Or your preferred Gemini model
instruction="You are a helpful assistant. Answer user questions using Google Search when needed.",
description="An assistant that can search the web.",
tools=[google_search]
)
Define a multi-agent system:
Define a multi-agent system with coordinator agent, greeter agent, and task execution agent. Then ADK engine and the model will guide the agents to work together to accomplish the task.
from google.adk.agents import LlmAgent, BaseAgent
# Define individual agents
greeter = LlmAgent(name="greeter", model="gemini-2.5-flash", ...)
task_executor = LlmAgent(name="task_executor", model="gemini-2.5-flash", ...)
# Create parent agent and assign children via sub_agents
coordinator = LlmAgent(
name="Coordinator",
model="gemini-2.5-flash",
description="I coordinate greetings and tasks.",
sub_agents=[ # Assign sub_agents here
greeter,
task_executor
]
)
Development UI
A built-in development UI to help you test, evaluate, debug, and showcase your agent(s).
Evaluate Agents
adk eval \
samples_for_testing/hello_world \
samples_for_testing/hello_world/hello_world_eval_set_001.evalset.json
🤝 Contributing
We welcome contributions from the community! Whether it's bug reports, feature requests, documentation improvements, or code contributions, please see our
- General contribution guideline and flow.
- Then if you want to contribute code, please read Code Contributing Guidelines to get started.
Community Repo
We have adk-python-community repo that is home to a growing ecosystem of community-contributed tools, third-party service integrations, and deployment scripts that extend the core capabilities of the ADK.
Vibe Coding
If you want to develop agent via vibe coding the llms.txt and the llms-full.txt can be used as context to LLM. While the former one is a summarized one and the later one has the full information in case your LLM has big enough context window.
Community Events
- [Completed] ADK's 1st community meeting on Wednesday, October 15, 2025. Remember to join our group to get access to the recording, and deck.
📄 License
This project is licensed under the Apache 2.0 License - see the LICENSE file for details.
Happy Agent Building!