Enforces grammar-based output constraints for LLM-generated text and structured data.
npx skills add https://github.com/lucasfth/config --skill guidance-lucasfth
Please help me install this Agent Skill. Skill: guidance Source: https://github.com/lucasfth/config/tree/main/.hermes/skills/mlops/inference/guidance Command: npx skills add https://github.com/lucasfth/config --skill guidance-lucasfth
Guidance enables precise control over LLM outputs using regex and grammar constraints to enforce structured formats (JSON/XML/code) and faster, deterministic results across workflows.
To get started, install Guidance and run a minimal constrained-generation example in Python.
High-intent search queries and answers about installing and using this skill.
Token healing fixes tokenization boundaries during constrained generation, preventing malformed outputs when regex or grammar constraints interrupt token sequences. It ensures grammar-driven generation produces clean, structured JSON or XML without broken syntax artifacts.
Yes, you can orchestrate complex multi-step workflows using Pythonic control flow to manage prompts across both local and cloud backends. This allows you to chain constrained generation steps dynamically for intricate form processing and data extraction pipelines.
The best way to enforce regex patterns is through grammar-driven generation, which intercepts the decoding process to only allow tokens matching your specified pattern. This guarantees outputs conform exactly to your regex constraints for reliable structured data.
Yes, the transformers library is required alongside the guidance package to apply grammar-based constraints and orchestrate constrained generation workflows. These dependencies provide the necessary backend integration for running local and cloud model inference with regex enforcement.
ihatesea69Enforce JSON, XML, and code output formats using Python.
Simon-Copilot-StudioEnforce valid JSON, XML, and code structures via grammar-based FSM constraints.
samuelmukotiControl LLM outputs with regex and grammars for structured JSON and XML formats.
ovachieverConstrain LLM outputs to valid JSON, XML, and code using regex and grammar constraints.
