What problem does it solve? When building structured AI pipelines, you often need to first understand what kind of text you are dealing with and then design an appropriate output format for it. Doing both steps ad hoc leads to inconsistent classifications and schemas that do not fit the input. This Skill chains the two steps into a measured, reproducible composition. ## Core Features & Use Cases - Text Classification: Categorizes any input text as a question, instruction, or narrative, with a one-sentence rationale and a confidence score written to classification.json. - Schema Design: Generates a JSON Schema (type, properties, required) tailored to both the input text and its classified category, written to schema.json. - Flexible Execution: Supports canonical per-step subagent dispatch or single-context inline execution when the skill is single-model. - Use Case: Given the text "Summarize the attached report in three bullet points and send it to the team", the skill classifies it as an instruction and produces a JSON Schema with fields for the action, target, recipients, and format constraints. ## Quick Start Classify this text and design a JSON output schema for a structured response to it: "What were the main causes of the supply shortage last quarter?"
