> ## Documentation Index > Fetch the complete documentation index at: https://praison.ai/docs/llms.txt > Use this file to discover all available pages before exploring further. # Task • AI Agent SDK > Task: A unit of work that can be executed by an Agent or a custom handler function. # Task > Defined in the [**task**](https://praison.ai/docs/sdk/reference/praisonaiagents/modules/task.md) module. AI Agent A unit of work that can be executed by an Agent or a custom handler function. Task is the unified abstraction for both AgentManager tasks and Workflow steps. It supports all features from the legacy Task class. Simple Usage: task = Task(description="Research AI trends") With action alias (from Task): task = Task(action="Write a blog post about \{\{topic}}") With custom handler function: task = Task( name="process\_data", action="Process the input", handler=my\_custom\_function ) With loop iteration: task = Task( action="Process \{\{item}}", loop\_over="items", loop\_var="item" ) ## Constructor **description** `Optional[str]` No description available. **expected\_output** `Optional[str]` No description available. **agent** `Optional[Agent]` No description available. **name** `Optional[str]` No description available. **tools** `Optional` No description available. **context** `Optional` No description available. **depends\_on** `Optional` No description available. **async\_execution** `Optional[bool]` (default: False) No description available. **config** `Optional` No description available. **output\_file** `Optional[str]` No description available. **output\_json** `Optional` No description available. **output\_pydantic** `Optional` No description available. **callback** `Optional` No description available. **on\_task\_complete** `Optional` No description available. **status** `str` (default: 'not started') No description available. **result** `Optional[TaskOutput]` No description available. **create\_directory** `Optional[bool]` (default: False) No description available. **id** `Optional[int]` No description available. **images** `Optional` No description available. **next\_tasks** `Optional` No description available. **task\_type** `str` (default: 'task') No description available. **human\_input** `bool` (default: False) No description available. **human\_review\_prompt** `Optional[str]` No description available. **condition** `Optional` No description available. **is\_start** `bool` (default: False) No description available. **loop\_state** `Optional` No description available. **memory** `Any` No description available. **quality\_check** `Any` (default: True) No description available. **input\_file** `Optional[str]` No description available. **rerun** `bool` (default: False) No description available. **retain\_full\_context** `bool` (default: False) No description available. **guardrail** `Optional` No description available. **guardrails** `Optional` No description available. **max\_retries** `int` (default: 3) No description available. **retry\_count** `int` (default: 0) No description available. **agent\_config** `Optional` No description available. **variables** `Optional` No description available. **skip\_on\_failure** `bool` (default: False) No description available. **retry\_delay** `float` (default: 0.0) No description available. **on\_error** `str` (default: 'stop') No description available. **action** `Optional[str]` No description available. **handler** `Optional[Callable]` No description available. **should\_run** `Optional[Callable]` No description available. **loop\_over** `Optional[str]` No description available. **loop\_var** `str` (default: 'item') No description available. **execution** `Optional` No description available. **routing** `Optional` No description available. **output\_config** `Optional` No description available. **output** `Optional` No description available. **when** `Optional[str]` No description available. **then\_task** `Optional[str]` No description available. **else\_task** `Optional[str]` No description available. **autonomy** `Optional` No description available. **knowledge** `Optional` No description available. **web** `Optional` No description available. **reflection** `Optional` No description available. **planning** `Optional` No description available. **hooks** `Optional` No description available. **caching** `Optional` No description available. **output\_variable** `Optional[str]` No description available. **fail\_on\_callback\_error** `bool` (default: False) No description available. **fail\_on\_memory\_error** `bool` (default: False) No description available. ## Methods > [card] [depends\_on()](https://praison.ai/docs/sdk/reference/praisonaiagents/functions/Task-depends_on.md) > > Alias for context - returns the list of dependent tasks. > [card] [depends\_on()](https://praison.ai/docs/sdk/reference/praisonaiagents/functions/Task-depends_on.md) > > Alias for context - sets the list of dependent tasks. > [card] [set\_status()](https://praison.ai/docs/sdk/reference/praisonaiagents/functions/Task-set_status.md) > > Set task status with transition validation. > [card] [initialize\_memory()](https://praison.ai/docs/sdk/reference/praisonaiagents/functions/Task-initialize_memory.md) > > Synchronous memory initialization with thread-safe locking > [card] [initialize\_memory\_async()](https://praison.ai/docs/sdk/reference/praisonaiagents/functions/Task-initialize_memory_async.md) > > Async-safe memory initialization with asyncio coordination > [card] [store\_in\_memory()](https://praison.ai/docs/sdk/reference/praisonaiagents/functions/Task-store_in_memory.md) > > Store content in memory with metadata > [card] [execute\_callback()](https://praison.ai/docs/sdk/reference/praisonaiagents/functions/Task-execute_callback.md) > > Execute callback and store quality metrics if enabled > [card] [execute\_callback\_sync()](https://praison.ai/docs/sdk/reference/praisonaiagents/functions/Task-execute_callback_sync.md) > > Synchronous wrapper that properly awaits execute\_callback instead of > [card] [evaluate\_when()](https://praison.ai/docs/sdk/reference/praisonaiagents/functions/Task-evaluate_when.md) > > Evaluate the 'when' condition against the given context. > [card] [get\_next\_task()](https://praison.ai/docs/sdk/reference/praisonaiagents/functions/Task-get_next_task.md) > > Get the next task name based on the 'when' condition evaluation. > [accordion] Internal & Generic Methods > > * **to\_dict**: Serialize the Task to a dictionary. ## Source > [card] [View on GitHub](https://github.com/MervinPraison/PraisonAI/blob/main/src/praisonai-agents/praisonaiagents/task/task.py#L21) > > `praisonaiagents/task/task.py` at line 21 *** ## Related Documentation > [card] [Tasks Concept](https://praison.ai/docs/docs/concepts/tasks) > [card] [Task Configuration](https://praison.ai/docs/docs/configuration/task-config) > [card] [Task Validation](https://praison.ai/docs/docs/features/task-validation-feedback) This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.