genai-toolbox

mirror of https://github.com/googleapis/genai-toolbox.git synced 2026-01-23 14:28:15 -05:00
Files
Mend Renovate cd4ae81501 chore(deps): update dependency google-adk to v1.18.0 (#1850 )
This PR contains the following updates:

| Package | Change | Age | Confidence |
|---|---|---|---|
| [google-adk](https://redirect.github.com/google/adk-python)
([changelog](https://redirect.github.com/google/adk-python/blob/main/CHANGELOG.md))
| `==1.15.1` -> `==1.18.0` |
[![age](https://developer.mend.io/api/mc/badges/age/pypi/google-adk/1.18.0?slim=true)](https://docs.renovatebot.com/merge-confidence/)
|
[![confidence](https://developer.mend.io/api/mc/badges/confidence/pypi/google-adk/1.15.1/1.18.0?slim=true)](https://docs.renovatebot.com/merge-confidence/)
|

---

### Release Notes

<details>
<summary>google/adk-python (google-adk)</summary>

###
[`v1.18.0`](https://redirect.github.com/google/adk-python/blob/HEAD/CHANGELOG.md#1180-2025-11-05)

[Compare
Source](https://redirect.github.com/google/adk-python/compare/v1.17.0...v1.18.0)

##### Features

- **\[ADK Visual Agent Builder]**
  - Core Features
    - Visual workflow designer for agent creation
- Support for multiple agent types (LLM, Sequential, Parallel, Loop,
Workflow)
    - Agent tool support with nested agent tools
    - Built-in and custom tool integration
- Callback management for all ADK callback types (before/after agent,
model, tool)
    - Assistant to help you build your agents with natural language
- Assistant proposes and writes agent configuration yaml files for you
    - Save to test with chat interfaces as normal
    - Build and debug at the same time in adk web!

- **\[Core]**
- Add support for extracting cache-related token counts from LiteLLM
usage
([4f85e86](4f85e86fc3))
- Expose the Python code run by the code interpreter in the logs
([a2c6a8a](a2c6a8a85c))
- Add run\_debug() helper method for quick agent experimentation
([0487eea](0487eea2ab))
- Allow injecting a custom Runner into `agent_to_a2a`
([156d235](156d235479))
- Support MCP prompts via the McpInstructionProvider class
([88032cf](88032cf5c5))

- **\[Models]**
- Add model tracking to LiteLlm and introduce a LiteLLM with fallbacks
demo
([d4c63fc](d4c63fc562))
- Add ApigeeLlm as a model that lets ADK Agent developers to connect
with an Apigee proxy
([87dcb3f](87dcb3f7ba))

- **\[Integrations]**
- Add example and fix for loading and upgrading old ADK session
databases
([338c3c8](338c3c89c6))
- Add support for specifying logging level for adk eval cli command
([b1ff85f](b1ff85fb23))
- Propagate LiteLLM finish\_reason to LlmResponse for use in callbacks
([71aa564](71aa5645f6))
- Allow LLM request to override the model used in the generate content
async method in LiteLLM
([ce8f674](ce8f674a28))
- Add api key argument to Vertex Session and Memory services for Express
Mode support
([9014a84](9014a849ea))
- Added support for enums as arguments for function tools
([240ef5b](240ef5beea))
- Implement artifact\_version related methods in GcsArtifactService
([e194ebb](e194ebb33c))

- **\[Services]**
- Add support for Vertex AI Express Mode when deploying to Agent Engine
([d4b2a8b](d4b2a8b49f))
- Remove custom polling logic for Vertex AI Session Service since LRO
polling is supported in express mode
([546c2a6](546c2a6816))
- Make VertexAiSessionService fully asynchronous
([f7e2a7a](f7e2a7a40e))

- **\[Tools]**
- Add Bigquery detect\_anomalies tool
([9851340](9851340ad1))
- Extend Bigquery detect\_anomalies tool to support future data anomaly
detection
([38ea749](38ea749c9c))
- Add get\_job\_info tool to BigQuery toolset
([6429457](64294572c1))

- **\[Evals]**
- Add "final\_session\_state" to the EvalCase data model
([2274c4f](2274c4f304))
- Marked expected\_invocation as optional field on evaluator interface
([b17c8f1](b17c8f19e5))
- Adds LLM-backed user simulator
([54c4ecc](54c4ecc733))

- **\[Observability]**
- Add BigQueryLoggingPlugin for event logging to BigQuery
([b7dbfed](b7dbfed4a3))

- **\[Live]**
- Add token usage to live events for bidi streaming
([6e5c0eb](6e5c0eb6e0))

##### Bug Fixes

- Reduce logging spam for MCP tools without authentication
([11571c3](11571c37ab))
- Fix typo in several files
([d2888a3](d2888a3766))
- Disable SetModelResponseTool workaround for Vertex AI Gemini 2+ models
([6a94af2](6a94af24bf))
- Bug when callback\_context\_invocation\_context is missing in
GlobalInstructionPlugin
([f81ebdb](f81ebdb622))
- Support models slash prefix in model name extraction
([8dff850](8dff85099d))
- Do not consider events with state delta and no content as final
response
([1ee93c8](1ee93c8bcb))
- Parameter filtering for CrewAI functions with \*\*kwargs
([74a3500](74a3500fc5))
- Do not treat FinishReason.STOP as error case for LLM responses
containing candidates with empty contents
([2f72ceb](2f72ceb49b))
- Fixes null check for reflect\_retry plugin sample
([86f0155](86f01550bd))
- Creates evalset directory on evalset create
([6c3882f](6c3882f2d6))
- Add ADK\_DISABLE\_LOAD\_DOTENV environment variable that disables
automatic loading of .env when running ADK cli, if set to true or 1
([15afbcd](15afbcd158))
- Allow tenacity 9.0.0
([ee8acc5](ee8acc58be))
- Output file uploading to artifact service should handle both base64
encoded and raw bytes
([496f8cd](496f8cd6bb))
- Correct message part ordering in A2A history
([5eca72f](5eca72f9bf))
- Change instruction insertion to respect tool call/response pairs
([1e6a9da](1e6a9daa63))
- DynamicPickleType to support MySQL dialect
([fc15c9a](fc15c9a0c3))
- Enable usage metadata in LiteLLM streaming
([f9569bb](f9569bbb1a))
- Fix issue with MCP tools throwing an error
([1a4261a](1a4261ad4b))
- Remove redundant `format` field from LiteLLM content objects
([489c39d](489c39db01))
- Update the contribution analysis tool to use original write mode
([54db3d4](54db3d4434))
- Fix agent evaluations detailed output rows wrapping
issue([4284c61](4284c61901))
- Update dependency version constraints to be based on PyPI
versions([0b1784e](0b1784e0e4))

##### Improvements

- Add Community Repo section to README
([432d30a](432d30af48))
- Undo adding MCP tools output schema to FunctionDeclaration
([92a7d19](92a7d19573))
- Refactor ADK README for clarity and consistency
([b0017ae](b0017aed44))
- Add support for reversed proxy in adk web
([a0df75b](a0df75b6fa))
- Avoid rendering empty columns as part of detailed results rendering of
eval results
([5cb35db](5cb35db921))
- Clear the behavior of disallow\_transfer\_to\_parent
([48ddd07](48ddd07894))
- Disable the scheduled execution for issue triage workflow
([a02f321](a02f321f1b))
- Include delimiter when matching events from parent nodes in content
processor
([b8a2b6c](b8a2b6c570))
- Improve Tau-bench ADK colab stability
([04dbc42](04dbc42e50))
- Implement ADK-based agent factory for Tau-bench
([c0c67c8](c0c67c8698))
- Add util to run ADK LLM Agent with simulation environment
([87f415a](87f415a7c3))
- Demonstrate CodeExecutor customization for environment setup
([8eeff35](8eeff35b35))
- Add sample agent for VertexAiCodeExecutor
([edfe553](edfe553942))
- Adds a new sample agent that demonstrates how to integrate PostgreSQL
databases using the Model Context Protocol (MCP)
([45a2168](45a2168e0e))
- Add example for using ADK with Fast MCP sampling
([d3796f9](d3796f9b33))
- Refactor gepa sample code and clean-up user demo
colab([63353b2](63353b2b74))

###
[`v1.17.0`](https://redirect.github.com/google/adk-python/blob/HEAD/CHANGELOG.md#1170-2025-10-22)

[Compare
Source](https://redirect.github.com/google/adk-python/compare/v1.16.0...v1.17.0)

##### Features

- **\[Core]**
- Add a service registry to provide a generic way to register custom
service implementations to be used in FastAPI server. See short
instruction
[here](https://redirect.github.com/google/adk-python/discussions/3175#discussioncomment-14745120).
([391628f](391628fcdc))
- Add the ability to rewind a session to before a previous invocation
([9dce06f](9dce06f9b0))
- Support resuming a parallel agent with multiple branches paused on
tool confirmation requests
([9939e0b](9939e0b087))
- Support content union as static instruction
([cc24d61](cc24d616f8))

- **\[Evals]**
- ADK cli allows developers to create an eval set and add an eval case
([ae139bb](ae139bb461))

- **\[Integrations]**
- Allow custom request and event converters in A2aAgentExecutor
([a17f3b2](a17f3b2e6d))

- **\[Observability]**
- Env variable for disabling llm\_request and llm\_response in spans
([e50f05a](e50f05a9fc))

- **\[Services]**
- Allow passing extra kwargs to create\_session of
VertexAiSessionService
([6a5eac0](6a5eac0bdc))
- Implement new methods in in-memory artifact service to support custom
metadata, artifact versions, etc.
([5a543c0](5a543c00df))
- Add create\_time and mime\_type to ArtifactVersion
([2c7a342](2c7a342593))
- Support returning all sessions when user id is none
([141318f](141318f775))

- **\[Tools]**
- Support additional headers for Google API toolset
([ed37e34](ed37e343f0))
- Introduces a new AgentEngineSandboxCodeExecutor class that supports
executing agent-generated code using the Vertex AI Code Execution
Sandbox API
([ee39a89](ee39a89110))
- Support dynamic per-request headers in MCPToolset
([6dcbb5a](6dcbb5aca6))
- Add `bypass_multi_tools_limit` option to GoogleSearchTool and
VertexAiSearchTool
([9a6b850](9a6b8507f0),
[6da7274](6da7274858))
- Extend `ReflectAndRetryToolPlugin` to support hallucinating function
calls
([f51380f](f51380f9ea))
- Add require\_confirmation param for MCP tool/toolset
([78e74b5](78e74b5bf2))

- **\[UI]**
- Granular per agent speech configuration
([409df13](409df1378f))

##### Bug Fixes

- Returns dict as result from McpTool to comply with BaseTool
expectations
([4df9263](4df926388b))
- Fixes the identity prompt to be one line
([7d5c6b9](7d5c6b9acf))
- Fix the broken langchain importing caused their 1.0.0 release
([c850da3](c850da3a07))
- Fix BuiltInCodeExecutor to support visualizations
([ce3418a](ce3418a69d))
- Relax runner app-name enforcement and improve agent origin inference
([dc4975d](dc4975dea9))
- Improve error message when adk web is run in wrong directory
([4a842c5](4a842c5a13))
- Handle App objects in eval and graph endpoints
([0b73a69](0b73a6937b))
- Exclude `additionalProperties` from Gemini schemas
([307896a](307896aece))
- Overall eval status should be NOT\_EVALUATED if no invocations were
evaluated
([9fbed0b](9fbed0b15a))
- Create context cache only when prefix matches with previous request
([9e0b1fb](9e0b1fb62b))
- Handle `App` instances returned by `agent_loader.load_agent`
([847df16](847df1638c))
- Add support for file URIs in LiteLLM content conversion
([85ed500](85ed500871))
- Only exclude scores that are None
([998264a](998264a5b1))
- Better handling the A2A streaming tasks
([bddc70b](bddc70b5d0))
- Correctly populate context\_id in remote\_a2a\_agent library
([2158b3c](2158b3c915))
- Remove unnecessary Aclosing
([2f4f561](2f4f5611bd))
- Fix pickle data was truncated error in database session using MySql
([36c96ec](36c96ec5b3))

##### Improvements

- Improve hint message in agent loader
([fe1fc75](fe1fc75c15))
- Fixes MCPToolset --> McpToolset in various places
([d4dc645](d4dc645478))
- Add span for context caching handling and new cache creation
([a2d9f13](a2d9f13fa1))
- Checks gemini version for `2 and above` for gemini-builtin tools
([0df6759](0df67599c0))
- Refactor and fix state management in the session service
([8b3ed05](8b3ed059c2))
- Update agent builder instructions and remove run command details
([89344da](89344da813))
- Clarify how to use adk built-in tool in instruction
([d22b8bf](d22b8bf890))
- Delegate the agent state reset logic to LoopAgent
([bb1ea74](bb1ea74924))
- Adjust the instruction about default model
([214986e](214986ebeb))
- Migrate invocation\_context to callback\_context
([e2072af](e2072af69f))
- Correct the callback signatures
([fa84bcb](fa84bcb575))
- Set default for `bypass_multi_tools_limit` to False for
GoogleSearchTool and VertexAiSearchTool
([6da7274](6da7274858))
- Add more clear instruction to the doc updater agent about one PR for
each recommended change
([b21d0a5](b21d0a50d6))
- Add a guideline to avoid content deletion
([16b030b](16b030b2b2))
- Add an sample agent for the `ReflectAndRetryToolPlugin`
([9b8a4aa](9b8a4aad6f))
- Improve error message when adk web is run in wrong directory
([4a842c5](4a842c5a13))
- Add an sample agent for the `ReflectAndRetryToolPlugin`
([9b8a4aa](9b8a4aad6f))
- Add span for context caching handling and new cache creation
([a2d9f13](a2d9f13fa1))
- Disable the scheduled execution for issue triage workflow
([bae2102](bae21027d9))
- Correct the callback signatures
([fa84bcb](fa84bcb575))

##### Documentation

- Format README.md for samples
([0bdba30](0bdba30263))
- Bump models in llms and llms-full to Gemini 2.5
([ce46386](ce4638651f))
- Update gemini\_llm\_connection.py - typo spelling correction
([e6e2767](e6e2767c39))
- Announce the first ADK Community Call in the README
([731bb90](731bb9078d))

###
[`v1.16.0`](https://redirect.github.com/google/adk-python/blob/HEAD/CHANGELOG.md#1160-2025-10-08)

[Compare
Source](https://redirect.github.com/google/adk-python/compare/v1.15.1...v1.16.0)

##### Features

- **\[Core]**
- Implementation of LLM context compaction
([e0dd06f](e0dd06ff04))
- Support pause and resume an invocation in ADK
([ce9c39f](ce9c39f5a8),

[2f1040f](2f1040f296),

[1ee01cc](1ee01cc05a),

[f005414](f005414895),

[fbf7576](fbf75761bb))
- **\[Models]**
- Add `citation_metadata` to `LlmResponse`
([3f28e30](3f28e30c6d))
- Add support for gemma model via gemini api
([2b5acb9](2b5acb98f5))
- **\[Tools]**
- Add `dry_run` functionality to BigQuery `execute_sql` tool
([960eda3](960eda3d1f))
- Add BigQuery analyze\_contribution tool
([4bb089d](4bb089d386))
- Spanner ADK toolset supports customizable template SQL and
parameterized SQL
([da62700](da62700d73))
- Support Oauth2 client credentials grant type
([5c6cdcd](5c6cdcd197))
- Add `ReflectRetryToolPlugin` to reflect from errors and retry with
different arguments when tool errors
([e55b894](e55b8946d6))
- Support using `VertexAiSearchTool` built-in tool with other tools in
the same agent
([4485379](4485379a04))
- Support using google search built-in tool with other tools in the same
agent
([d3148da](d3148dacc9))
- **\[Evals]**
- Add HallucinationsV1 evaluation metric
([8c73d29](8c73d29c75))
- Add Rubric based tool use metric
([c984b9e](c984b9e552))
- **\[UI]**
- Adds `adk web` options for custom logo
([822efe0](822efe0065))
- **\[Observability]**
- **otel:** Switch CloudTraceSpanExporter to telemetry.googleapis.com
([bd76b46](bd76b46ce2))

##### Bug Fixes

- Adapt to new computer use tool name in genai sdk 1.41.0
([c6dd444](c6dd444fc9))
- Add AuthConfig json serialization in vertex ai session service
([636def3](636def3687))
- Added more agent instructions for doc content changes
([7459962](745996212d))
- Convert argument to pydantic model when tool declares it accepts
pydantic model as argument
([571c802](571c802fba))
- Do not re-create `App` object when loader returns an `App`
([d5c46e4](d5c46e4960))
- Fix compaction logic
([3f2b457](3f2b457efd))
- Fix the instruction in workflow\_triage example agent
([8f3ca03](8f3ca0359e))
- Fixes a bug that causes intermittent `pydantic` validation errors when
uploading files
([e680063](e68006386f))
- Handle A2A Task Status Update Event when streaming in
remote\_a2a\_agent
([a5cf80b](a5cf80b952))
- Make compactor optional in Events Compaction Config and add a default
([3f4bd67](3f4bd67b49))
- Rename SlidingWindowCompactor to LlmEventSummarizer and refine its
docstring
([f1abdb1](f1abdb1938))
- Rollback compaction handling from \_get\_contents
([84f2f41](84f2f417f7))
- Set `max_output_tokens` for the agent builder
([2e2d61b](2e2d61b6fe))
- Set default response modality to AUDIO in run\_session
([68402bd](68402bda49))
- Update remote\_a2a\_agent to better handle streaming events and avoid
duplicate responses
([8e5f361](8e5f361264))
- Update the load\_artifacts tool so that the model can reliably call it
for follow up questions about the same artifact
([238472d](238472d083))
- Fix VertexAiSessionService base\_url override to preserve initialized
http\_options
([8110e41](8110e41b36),
[c51ea0b](c51ea0b52e))
- Handle `App` instances returned by `agent_loader.load_agent`
([847df16](847df1638c))

##### Improvements

- Migrate VertexAiSessionService to use Agent Engine SDK
([90d4c19](90d4c19c51))
- Migrate VertexAiMemoryBankService to use Agent Engine SDK
([d1efc84](d1efc8461e),
[97b950b](97b950b36b),
[83fd045](83fd045718))
- Add support for resolving $ref and $defs in OpenAPI schemas
([a239716](a239716930))

##### Documentation

- Update BigQuery samples README
([3021266](30212669ff))

</details>

---

### Configuration

📅 **Schedule**: Branch creation - At any time (no schedule defined),
Automerge - At any time (no schedule defined).

🚦 **Automerge**: Disabled by config. Please merge this manually once you
are satisfied.

♻ **Rebasing**: Whenever PR becomes conflicted, or you tick the
rebase/retry checkbox.

🔕 **Ignore**: Close this PR and you won't be reminded about this update
again.

---

- [ ] <!-- rebase-check -->If you want to rebase/retry this PR, check
this box

---

This PR was generated by [Mend Renovate](https://mend.io/renovate/).
View the [repository job
log](https://developer.mend.io/github/googleapis/genai-toolbox).

<!--renovate-debug:eyJjcmVhdGVkSW5WZXIiOiI0MS4xNTkuNCIsInVwZGF0ZWRJblZlciI6IjQxLjE1OS40IiwidGFyZ2V0QnJhbmNoIjoibWFpbiIsImxhYmVscyI6W119-->

Co-authored-by: Averi Kitsch <akitsch@google.com>