Parse LLM output as AgentResponse or return raw agent response.

Prepare generation config for Google Gemini API. Args: system_instruction: System instruction for the model tools: To

Prepare common setup for kickoff execution. This method handles all the common preparation logic shared between kickoff() and kickoff

Resolve an HTTPS MCP server URL into tools.

Resolve an ``MCPServerConfig`` into tools. Returns ``(tools, clients)`` where *clients* is always empty for native tools (clients are

Build a ``_ScriptedFlow``, attach ``script``, kickoff. Returns the flow.

Path-traversal-safe extraction for Python versions without tar filters. Validates both the member's own path and, for symlink/hardlink members,

Save task output to a file. Note: For cross-platform file writing, especially on Windows, consider using FileWriterTool

Scan rows optionally filtered by scope prefix. Uses a full table scan (no vector query) so the limit is applied after the scope filte

Display a standard error message when no UUID or project name is available.

Unregister a stream handler from the event bus. Args: handler: The handler function to unregister.

Validate parameters before making a call. Currently this only checks if a response_format is provided and whether the model supports

Validate file size against maximum. Args: file_type: Type label for error messages (e.g., "Image", "PDF"). filename: Name of the

Create a new collection in Qdrant asynchronously. Keyword Args: collection_name: Name of the collection to create. Must be unique

Download a template and copy it into the current working directory. Args: name: Template name (with or without the template_ pref

Add usage metrics from another UsageMetrics object. Args: usage_metrics: The usage metrics to add.

Delete a collection and all its data asynchronously. Keyword Args: collection_name: Name of the collection to delete. Ra

Execute a task with the agent asynchronously. Args: task: Task to execute. context: Context to execute the task in.

Asynchronously invoke the tool. Args: input: The input arguments config: Optional configuration **kwargs:

Async read the file content as bytes. Raises: TypeError: If the underlying source doesn't support async read.

Run the flow asynchronously and stream scoped public frames.

Check if max iterations reached before proceeding with reasoning.

Write a checkpoint. Args: location: Storage destination. For JsonProvider this is a directory path; for SqlitePro

Write a JSON checkpoint file. Args: data: The serialized JSON string to persist. location: Base directory where check

Classify an exception as transient or permanent upload error. Analyzes the exception type name and status code to determine if the error is l

Sync wrapper for aclear.

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

Clear files for a task execution. Args: task_id: Unique identifier for the task.

Return a configured CrewAI definition path for a project type. ``[tool.crewai].type`` must match ``project_type`` and ``definition`` must be

Convert a path to a Path object.

Create a converter instance based on the agent or provided class. Args: agent: The agent instance. converter_cls: The converter c

Create ``branch`` from current HEAD, resetting any stale copy. If the branch exists locally or on origin, prompts the user to choose between

Disconnect from MCP server and cleanup resources.

Generate embedding for a single text. Args: text: Text to embed Returns: List of floats representin

Ensure all schema objects in anyOf/oneOf have a 'type' key. OpenAI strict mode requires every schema to have a 'type' key. Empty schemas {} i

Execute a tool and check if the result should be treated as a final answer. This function integrates tool hooks for before and after tool executi

Extract ``(finish_reason, response_id)`` from a choices-shaped response. Handles both object-style (``response.id``, ``response.choices[0].finish

Extract the checkpoint ID from a location string. Args: location: The identifier returned by a previous ``checkpoint`` call.

Extract knowledge from the task prompt.

Extract discrete memories from content. Delegates to this flow's memory. Args: content: Raw text (e.g. task + result dump).

Extract tool call ID, name, and arguments from various provider formats. Args: tool_call: The tool call object to extract info from.

Format an error message with the given template and error. Args: template: The error message template to use error: T

Sync wrapper for aget_all_for_provider.

Get all completed todos. Returns: List of completed TodoItem objects.

Returns the context window size, using 75% of the maximum to avoid cutting off messages mid-thread. Raises: Valu

Get the ID of the last emitted event for linear chain tracking. Returns: The event_id of the previously emitted event, or None if no even

Get an existing collection or create it if it doesn't exist. Returns existing collection if found, otherwise creates a new one. Keyw

Get the page count of a PDF file. Args: file: The PDF file to measure. Returns: Number of pages, or None if page count canno

Get all todos that are ready to execute (pending with satisfied dependencies). Returns: List of TodoItem objects that can be exec

Get the sanitized names of the tools. Args: tools: List of tools to get names from. Returns: Comma-separated string of sanit

Show method status panel.

Return True if the Git repository is fully synced with the remote, False otherwise.

Execute the agent with the given messages. Args: messages: Either a string query or a list of message dictionaries.

Start the flow execution asynchronously. This method performs state restoration (if an 'id' is provided and persistence is available)

Acquire a named lock, yielding while it is held. Args: name: A human-readable lock name (e.g. ``"chromadb_init"``). The bui

Mark that events have been collected during execution.

Record a tool usage event in the current trace. Args: tool_name: Name of the tool used. tool_args: Arguments passed t

Optimize an image to fit within a target file size. Uses iterative quality reduction to achieve target size. Args: file: The image f

Process the LLM response and format it into an AgentAction or AgentFinish. Args: answer: The raw response from the LLM use_stop_w

Process response through all registered extensions. Args: agent_response: The parsed agent response. extension_states

Process A2A task state and return result dictionary. Shared logic for both polling and streaming handlers. Args: a2a_task: The A2A t

Interpolate ``${...}`` expressions inside nested strings as CEL. A string that is exactly one ``${...}`` keeps the evaluated value's

Request user input via console (blocking). Displays the prompt message with formatting and waits for the user to type their response.

Clear all stored task outputs.

Resolve a registry reference to a Skill object. Resolution order: 1. ./skills/{name}/ in the current working directory (project-local) 2.

Set reference to batch manager for sending events. Args: batch_manager: The trace batch manager instance.

Set whether to suppress tracing-related console messages. Args: suppress: True to suppress messages, False to show them. Returns:

Determine if tracing should be enabled. This is the single source of truth for tracing enablement. Priority order: 1. Explicit override (

Check if tracing messages should be suppressed. Returns: True if messages should be suppressed, False otherwise.

Store files for a task execution. Args: task_id: Unique identifier for the task. files: Dictionary mapping names to file inputs.

Remove null type from anyOf/type arrays. Pydantic generates `T | None` for optional fields, which creates schemas with null in the type. Howe

Serialize context to a dictionary for persistence. Returns: Dictionary representation suitable for JSON serialization.

Fetch actions and return built tools.