exo:models

Configure LLM providers and models for Exo agents via ModelConfig.

1|Updated Feb 16, 2026

Search Tags:#llm #configure #provider #token-count #context-windows #model-config #get_provider

Ready to run

npx skills add https://github.com/midsphere-ai/exo --skill exo-models

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: exo:models
Source: https://github.com/midsphere-ai/exo/tree/main/skills/models
Command: npx skills add https://github.com/midsphere-ai/exo --skill exo-models

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Configures which LLM provider and model an Exo agent uses, including API keys, endpoints, and custom providers, enabling seamless model swapping.

Core Features & Use Cases

Model string parsing and provider selection (provider:model) with lazy provider resolution
Manage ModelConfig, context windows, and token counting across multiple providers
Register custom providers and integrate multimodal tools via provider-specific options

Quick Start

Create an Agent with model "openai:gpt-4o" and ensure the corresponding API key is available in your environment.

Frequently Asked Questions about exo:models

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

FAQPage Schema

How do I configure LLM provider and model settings for an Exo agent?▼

To configure LLM provider settings, create an Agent with a model string like "openai:gpt-4o" and ensure the corresponding API key is available in your environment variables for seamless model swapping.

Can I use custom providers and context windows with OpenAI or Anthropic models?▼

Yes, you can register custom providers and integrate optional context window mappings to tailor provider behavior across OpenAI, Anthropic, Gemini, and Vertex for multimodal tools.

How does model string parsing work for provider selection?▼

Model string parsing uses the provider:model format to enable lazy provider resolution, allowing the Exo agent to dynamically select the correct LLM provider and model configuration.

Do I need environment variables to manage API keys and endpoints?▼

Yes, environment variables are required to securely supply API keys and endpoints, which ModelConfig and the provider registry use to authenticate and manage your selected LLM providers.

What is the best way to handle token counting across multiple LLM providers?▼

Handling token counting across multiple providers requires using ModelConfig alongside the model_registry, ensuring accurate context window mappings and provider-specific options for multimodal tools.