goal_parsing

To extract missing slots and build a structured JSON state, this skill guides users through a deterministic decision-tree interview, identifying required fields like user_type and scenario. It returns a portrait JSON containing missing_slots and next_questions to drive subsequent planning agents.

Slot-filling for dialogue management is the process of navigating a decision tree to collect required target parameters. You need it when shaping multi-faceted goals for downstream planning workflows, ensuring all core fields are captured before executing a provisioning task.

You can track conversation state and determine the next question by feeding composite goals to this skill. It maintains and updates a JSON state across interactions, outputting an is_complete flag and next_questions to seamlessly drive the subsequent steps of your planning agent.

Yes, the slot-filling skill supports YAML configurations for state tracking, operating with the pyyaml dependency. It processes natural language input alongside your defined slots to output a structured JSON portrait without requiring complex environmental setups.

The best way to collect a complete target portrait is using a deterministic slot-filling workflow. This approach navigates a defined sequence to capture seven core slots, returning a structured JSON state that downstream planning and provisioning agents can directly consume.

When required slots like time_window or package_type are missing during goal parsing, the output JSON explicitly flags them in the missing_slots array. The skill then generates specific next_questions to prompt the user, ensuring the target portrait reaches is_complete status.