CREATE AN ML-BASED ACTION

Creates a new ML-based action that links a trained machine learning model to a project and defines when the action is triggered and whether human inspection is required. Use this worker to register automated ML-driven decision actions within the platform’s assembler pipeline.

When to use

Tagged: action_trigger, assembler, automation, human_in_the_loop, ml_action.

Inputs

Label	ID	Type	Default	Required	Description
Name	name	text	—	✓	Human-readable display name for the ML action; used to identify the action in the platform UI and audit logs.
Action Type	type	text	—	✓	Categorical identifier for the action type (e.g., ‘classification’, ‘regression’, ‘anomaly’); determines how the ML model output is interpreted and acted upon.
When to take Action	use_on	text	—	✓	Trigger condition or event that activates this action (e.g., ‘on_simulation_complete’, ‘on_file_upload’); controls when the linked ML model is invoked.
Human in the loop	user_inspection	select	—	✓	Whether a human reviewer must approve the ML model’s output before the action is executed; select ‘yes’ to enable human-in-the-loop gating, ‘no’ for fully automated execution.
Machine-Learning Model	mathmodel_id	remote_lookup	—	✓	Reference to the trained machine learning model (math model) that powers this action; selected via remote lookup from the platform’s registered ML model catalogue.
Project	project_id	remote_lookup	—	✓	The platform project to which this ML action is scoped; selected via remote lookup and restricts the action’s visibility and data access to that project.

Outputs

Label	ID	Type	Description
ML Action ID	mlaction_id	integer	Integer primary key of the newly created ML action record; use this ID to reference, update, or trigger the action in downstream workflow steps.

Disciplines

• ai_ml.agents
• platform.integration
• platform.workflow

---

Auto-generated from platform schema. Worker id: assembler_create_mlaction. Schema hash: e9b598817bae. Hand-curated docs in workerexamples/ override this page when present.