[mohitdixit02]

Link to Issue or Description of Change

1. Link to an existing issue (if applicable):

• Closes: LlmAgent(mode="task") with a raw dict output_schema fails at construction with SchemaError: Unknown schema type: "object" #7935

Problem:
When creating an LlmAgent in task mode (mode="task") with a raw JSON schema dictionary (or a google.genai.types.Schema instance) as its output_schema, agent initialization failed at construction with:
pydantic_core._pydantic_core.SchemaError: Unknown schema type: "object" (or PydanticSchemaGenerationError).

In FinishTaskTool.__init__, self._adapter: TypeAdapter[Any] = TypeAdapter(self.output_schema) assumed output_schema was always a Python type or Pydantic
model. Because TypeAdapter does not accept raw JSON schema dictionaries or types.Schema instances, construction failed before the agent could be used.

Solution:

1. In FinishTaskTool.__init__:
   • Set self._adapter = None when output_schema is a raw dict or types.Schema, bypassing Pydantic's TypeAdapter.
   • Pass self.output_schema directly to get_output_wrapper_key.
2. In FinishTaskTool._get_declaration():
   • When self._adapter is None, obtain the raw schema via copy.deepcopy(schema_to_json_schema(self.output_schema)) to prevent in-place mutation of the user's schema dictionary.
3. In FinishTaskTool.run_async():
   • When self._adapter is None, pass raw_value through directly, matching the established precedent in SetModelResponseTool and validate_node_data.
4. In schema_to_json_schema():
   • Added support for types.Schema using model_dump(mode="json", exclude_none=True) and lowercase_schema_types.

Testing Plan

Unit Tests:

• I have added or updated unit tests for my change.
• All unit tests pass locally.

Summary of passed pytest results:

pytest tests/unittests/agents/llm/task/test_finish_task_tool.py tests/unittests/utils/test_schema_utils.py

collected 136 items
tests/unittests/agents/llm/task/test_finish_task_tool.py ..................................................... [ 38%]
tests/unittests/utils/test_schema_utils.py ......................................................             [100%]
====================== 136 passed in 1.69s ======================

Manual End-to-End (E2E) Tests:
Verified against the reproduction case with a raw dictionary schema:

    schema = {"type": "object", "properties": {"answer": {"type": "string"}}}
    agent = LlmAgent(name="b", model="gemini-2.5-flash", output_schema=schema, mode="task")

• Before: Raised pydantic_core._pydantic_core.SchemaError: Unknown schema type: "object".
• After: Constructs cleanly and registers finish_task tool with unwrapped parameters declaration.

Checklist

• I have read the CONTRIBUTING.md document.
• I have performed a self-review of my own code.
• I have commented my code, particularly in hard-to-understand areas.
• I have added tests that prove my fix is effective or that my feature works.
• New and existing unit tests pass locally with my changes.
• I have manually tested my changes end-to-end.
• Any dependent changes have been merged and published in downstream modules.