Resolve kickoff inputs for a declarative crew. Mirrors the declarative-flow experience (``_resolve_flow_inputs``): layers ``--inputs`` over t

Resolve crew-level skill paths once so agents don't repeat the work.

Resolve a potentially async result to its value.

Execute the Agent Handler tool with the given arguments.

Send buffered events to backend with graceful failure handling

Set fingerprint data on an event from an agent object.

Update a flow method step's status. Caller must hold ``self._lock``.

Set the final result after streaming completes. Args: result: The final CrewOutput.

Set the read and write streams. Args: read: Read stream. write: Write stream.

Setup and invoke after_llm_call hooks for the executor context. Args: executor_context: The executor context to setup the hooks for.

True when the active trace belongs to a deferred flow session.

Reject like GPT-5.6 until an explicit reasoning_effort="none" arrives. The reasoning planner calls the LLM with its own `available_functions`

Return a stable lock name for the user data file.

Validate content type against supported formats. Args: file_type: Type label for error messages (e.g., "Image", "Audio"). filenam

Async call to Anthropic messages API. Args: messages: Input messages for the chat completion tools: List of tool/func

Clear files for a crew execution asynchronously. Args: execution_id: Unique identifier for the crew execution.

Add an edge from this node to another. Args: edge_type: The relationship type. target_id: The event_id of the target

Async delete an uploaded file. Default implementation runs sync delete in executor. Override in subclasses for native async support.

Get a cached upload by file hash. Args: file_hash: Hash of the file content. provider: The provider name. Re

Retrieve files for a crew execution asynchronously. Args: execution_id: Unique identifier for the crew execution. Returns: D

Record an agent result, optionally making it visible to the user.

Append a message to ``state.messages`` or the flow fallback buffer.

Write an ExecutionContext back into the ContextVars.

Save documents to the knowledge base asynchronously. Args: documents: List of document strings to save.

Search for documents in the knowledge base asynchronously. Args: query: List of query strings. limit: Maximum number

Cache an uploaded file by hash. Args: file_hash: Hash of the file content. content_type: MIME type of the file.

Store files for a crew execution asynchronously. Args: execution_id: Unique identifier for the crew execution. files: Dictionary

Rebind the runtime ``Memory`` dependency after restore. Required after deserializing from a checkpoint, since the live ``Memory`` can

Return a copy of ``os.environ`` augmented with UV_INDEX_* credentials for ``repository_handle``. The handle is normalized to upper-case with

Build an execution plan from handlers and their dependencies. Args: handlers: All handlers for an event type dependencies: Mappin

Return the bytes of a one-page PDF containing *text*.

Build instructions for downgrading native tools to text tool calls.

Calculates and returns the usage metrics.

Call Azure AI Inference API. Args: messages: Input messages tools: List of tool/function definitions call

Mock LLM call that returns a predefined response. Properly formats messages to match OpenAI's expected structure.

Execute LLM call with hooks and parse the response. Returns routing decision based on parsing result.

Override the stop list for ``llm`` within the current call scope. Only ``llm``'s reads via :attr:`BaseLLM.stop_sequences` see ``stop``; other

Check if an output type is set.

Chunk the input text into smaller pieces. Args: text: The text to chunk. Returns: A list of text chunks.

Map user text to one of the given outcomes using an LLM.

Close the underlying stream.

Records which AI coding assistant (if any) is running this process. Emitted at most once per process as a feature usage event, so it lands

Compute a collision-resistant key using SHA-256.

Compute the SHA-256 hash of the given content. Args: content: The content to hash. Returns: The SHA-256 hash of the content as a

Configure HTTP client with auth-specific settings. Only HTTPDigestAuth and APIKeyAuth need client configuration. Args: auth: Authent

Copy a file from src to dst.

Create an async chunk generator that uses a holder to access streaming output. Args: state: The streaming state. run_coro: Corout

Create a new crew deployment. Args: confirm (bool): Whether to skip the interactive confirmation prompt. ski

Create a scoped synchronous public frame generator.

Create Recall and Remember tools for the given memory instance. When memory is read-only (``_read_only=True``), only the RecallMemoryTool is

Create a loader tool when metadata-only skills are available. Args: skills: Skills configured on the agent. source: Event source

Create and register streaming state. Args: current_task_info: Task context info. result_holder: List to hold the final result.

Return the per-call usage accrued since ``baseline`` was captured. Both objects must come from the same monotonically increasing accu

Return all descendant nodes, children recursively. Args: event_id: The root event ID to start from. Returns:

Generate embeddings for multiple texts. Args: texts: List of texts to embed Returns: List of embedd

Embed multiple texts in a single API call. The embedder already accepts ``list[str]``, so this just calls it once with the full batch and nor

Enable streaming on all agents that have an LLM configured. Args: agents: Iterable of agents to enable streaming on.