airflow.providers.common.ai.decorators.agent¶

TaskFlow decorator for agentic LLM workflows.

The user writes a function that returns the prompt string. The decorator handles hook creation, agent configuration with toolsets, multi-turn reasoning, and output serialization.

Functions¶

agent_task([python_callable])

Wrap a function that returns a prompt into an agentic LLM task.

Module Contents¶

airflow.providers.common.ai.decorators.agent.agent_task(python_callable=None, **kwargs)[source]¶

Wrap a function that returns a prompt into an agentic LLM task.

The function body constructs the prompt (can use Airflow context, XCom, etc.). The decorator handles hook creation, agent configuration with toolsets, multi-turn reasoning, and output serialization.

Usage:

@task.agent(
    llm_conn_id="pydanticai_default",
    system_prompt="You are a data analyst.",
    toolsets=[SQLToolset(db_conn_id="postgres_default")],
)
def analyze(question: str):
    return f"Answer: {question}"

Parameters:: python_callable (collections.abc.Callable | None) – Function to decorate.