fix(agents): keep null in the task output schema embedded in the prompt (#6775)

* fix(agents): keep null in the task output schema embedded in the prompt

build_task_prompt_with_schema embeds the task output schema into the prompt
via generate_model_description, whose strip_null_types defaults to True.
Combined with ensure_all_properties_required, an Optional[str] = None field
reaches the model as a required, non-nullable string, contradicting the
provider-side response schema generated from the same model.

That sanitizer targets OpenAI strict function-calling schemas. This call site
produces prompt prose, where those constraints do not apply.

Pass strip_null_types=False, matching the existing call for tool schemas in
utilities/agent_utils.py.

Fixes #6774

* test(agents): cover the output_json branch of the prompt schema

build_task_prompt_with_schema embeds a schema on both the output_json and
the output_pydantic branch, and this PR changes both. The regression test
only built a Task with output_pydantic, so the output_json branch shipped
unpinned.

Parameterize over both output attributes. Checked on this branch with the
fix reverted: both cases fail on the missing anyOf, and both pass with it.

Also compare the anyOf members as a set, so member order is not part of the
test contract.

---------

Co-authored-by: Vidit Ostwal <110953813+Vidit-Ostwal@users.noreply.github.com>

lib/crewai/src/crewai/agent/utils.py

@@ -74,13 +74,17 @@ def build_task_prompt_with_schema(task: Task, task_prompt: str) -> str:
 
     if (task.output_json or task.output_pydantic) and not task.response_model:
         if task.output_json:
-            schema_dict = generate_model_description(task.output_json)
+            schema_dict = generate_model_description(
+                task.output_json, strip_null_types=False
+            )
             schema = json.dumps(schema_dict["json_schema"]["schema"], indent=2)
             task_prompt += "\n" + I18N_DEFAULT.slice(
                 "formatted_task_instructions"
             ).format(output_format=schema)
         elif task.output_pydantic:
-            schema_dict = generate_model_description(task.output_pydantic)
+            schema_dict = generate_model_description(
+                task.output_pydantic, strip_null_types=False
+            )
             schema = json.dumps(schema_dict["json_schema"]["schema"], indent=2)
             task_prompt += "\n" + I18N_DEFAULT.slice(
                 "formatted_task_instructions"

lib/crewai/tests/agents/test_agent_utils.py

@@ -0,0 +1,41 @@
+"""Tests for crewai.agent.utils prompt-building helpers."""
+
+from __future__ import annotations
+
+import json
+
+import pytest
+from pydantic import BaseModel, Field
+
+from crewai import Task
+from crewai.agent.utils import build_task_prompt_with_schema
+
+
+class _Output(BaseModel):
+    name: str
+    note: str | None = Field(default=None)
+
+
+@pytest.mark.parametrize("output_attribute", ["output_pydantic", "output_json"])
+def test_optional_fields_stay_nullable_in_the_prompt_schema(
+    output_attribute: str,
+) -> None:
+    """An Optional field must still be expressible as null in the prompt schema.
+
+    The provider-side response schema generated from the same model allows null,
+    so stripping it here hands the model two contradictory contracts and leaves
+    it no way to say "not applicable". Both output attributes embed a schema in
+    the prompt, so both are pinned.
+    """
+    task = Task(description="d", expected_output="e", **{output_attribute: _Output})
+
+    prompt = build_task_prompt_with_schema(task, "")
+
+    start = prompt.index("{")
+    end = prompt.rindex("}", start) + 1
+    schema = json.loads(prompt[start:end])
+
+    assert {entry["type"] for entry in schema["properties"]["note"]["anyOf"]} == {
+        "string",
+        "null",
+    }