What problem does it solve? LLMs frequently produce malformed JSON, invalid dates, or off-format text that breaks downstream parsing and forces retry loops. This Skill uses Microsoft Research's Guidance framework to constrain generation at the token level, so outputs always match the required regex, grammar, or selection set. ## Core Features & Use Cases - Regex and Grammar Constraints: Enforce formats like emails, dates, phone numbers, UUIDs, and full JSON/XML structures during generation. - Token Healing: Automatically fixes tokenization boundaries between prompts and generated text for natural continuations. - Multi-Step Workflows: Build ReAct agents, chain-of-thought reasoning, and classification pipelines with Pythonic control flow and reusable @guidance functions. - Use Case: Extract structured entities (person, organization, date, location) from unstructured text with guaranteed format compliance, or generate schema-valid JSON for API payloads without validation retries. ## Quick Start Use the guidance skill to generate a JSON user profile with constrained name, age, and email fields using Claude.
