Tool Node

Provides a mechanism for directly and deterministically executing a specific, pre-defined Lyntaris Tool within the workflow sequence. Unlike the Agent node, where the LLM dynamically chooses a tool based on reasoning, the Tool node executes exactly the tool selected by the workflow designer.
Functionality
This node is used when the workflow rigorously requires the explicit execution of a known capability at a defined point, given deterministically available inputs. It completely negates the hallucination risks of involving an LLM reasoning engine for tool selection while passing context variables safely into execution parameters.
Execution Sequence
- Triggering: The runtime execution thread arrives at the Tool node.
- Identification: Identifies the precise Lyntaris C# / Python mapped tool selected in its configuration.
- Resolution: Evaluates required tool mapping properties, interpolating dynamic
$flow.stateinjections. - Execution: Dispatches the execution request securely to backend subprocesses.
- Propagation: Casts the serialized result directly to the output socket.
Configuration Parameters
- Tool Selection: Choose the specifically registered Lyntaris Tool from a dynamically fetched platform dropdown.
- Input Arguments: Dynamically adapts based on the tool's expected schema constraints.
- Map Argument Name: Maps inputs to tool properties natively.
- Provide Argument Value: Supports raw strings, or
{{ previousNode.output }},{{ $flow.state.data }}references.
- Update Flow State: Bind extracted tool payload logic securely directly into
$flow.statekeys for subsequent pipeline usages without parsing overhead.
Inputs & Outputs
- Inputs: Receives necessary data for arguments sourcing values dynamically.
- Outputs: Produces the raw executed output (e.g., JSON response, raw numbers).