skills.rest
ChrstprJohn
ChrstprJohnCOMMUNITY
Explore all skills by this author →
GitHub

llm-structured-output

Extract typed, validated data from LLM responses using provider-specific schema strategies.

6|Updated Mar 1, 2026
Search Tags:#llm#openai#anthropic#structured-output#json-schema#constrained-decoding#google-gemini
One-click install
npx skills add https://github.com/ChrstprJohn/SamsonDentalCenter --skill llm-structured-output
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: llm-structured-output
Source: https://github.com/ChrstprJohn/SamsonDentalCenter/tree/main/.agent/skills/llm-structured-output
Command: npx skills add https://github.com/ChrstprJohn/SamsonDentalCenter --skill llm-structured-output

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Replaces ad-hoc parsing of LLM responses with typed, validated data by extracting JSON-like structures from model outputs.

Core Features & Use Cases

  • Structured extraction across OpenAI, Anthropic, and Google Gemini APIs using provider-specific strategies (response_format with json_schema, tool_use blocks, and responseSchema) to guarantee schema conformance.
  • Schema-driven validation and retry logic to handle decoding failures, malformed JSON, or missing fields in production systems.
  • Supports defining target schemas for downstream pipelines, databases, or UI layers, with step-by-step guidance on choosing the right method per provider.

Quick Start

Create a concrete target schema, configure the extraction provider, and run a single-pass extraction on your LLM output to obtain a typed data object.

Frequently Asked Questions about llm-structured-output

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I get structured JSON output from LLM responses instead of parsing free text?▼

To get structured JSON output from LLM responses, you define a target schema and apply provider-specific strategies like OpenAI's response_format with json_schema, Anthropic's tool_use blocks, or Google Gemini's responseSchema to guarantee schema conformance.

What's the best way to extract typed data from OpenAI, Anthropic, and Google Gemini models?▼

The best way to extract typed data is using provider-specific extraction strategies: OpenAI's response_format with json_schema, Anthropic's tool_use blocks, and Google Gemini's responseSchema, ensuring deterministic extraction and schema validation across different APIs.

Why does my LLM return invalid or incomplete JSON when I need structured data?▼

Invalid or incomplete JSON occurs when parsing free-text without validation. Implement schema-driven validation and retry logic to handle decoding failures, malformed JSON, or missing fields, guarding your production systems against invalid LLM outputs.

Can I use a single schema definition for structured extraction across multiple LLM providers?▼

Yes, you can define a concrete target schema for downstream pipelines, databases, or UI layers. The extraction process applies provider-specific strategies to guarantee schema conformance, allowing consistent typed data extraction across OpenAI, Anthropic, and Google Gemini.

How do I validate LLM structured outputs to handle malformed JSON in production?▼

To validate LLM structured outputs and handle malformed JSON in production, implement deterministic extraction followed by schema-driven validation and retry logic. This guards downstream pipelines against invalid or incomplete data from model responses.

Related Skills

Explore More Software Engineering Skills→
SuperlendSuperlend

libafl

Build custom fuzzers with modular Rust components and pluggable strategies.

Official
Advanced
JigsawStackJigsawStack
4

structured-output

Convert prompts into schema-constrained JSON objects using Zod or Pydantic.

Official
Intermediate
hpsgdhpsgd

write-schema

Generate JSON Schema and Pydantic models for Python configuration validation.

Official
Advanced
vecearvecear

structured-output-extractor

Extract typed JSON data from unstructured LLM outputs using Zod validation.

Community
Advanced