model-discovery

To fetch current AI models without hardcoding, use a fetch/cache mechanism that retrieves up-to-date model names from providers and checks them against a centralized known-model registry.

Classify AI models into fast, default, and heavy tiers by applying a tier-mapping configuration to the fetched model names, which guides categorization for routing and usage across AI services.

Yes, you can automatically detect new AI models from Anthropic, OpenAI, Gemini, and Ollama by running a workflow that compares live provider data against a known-model registry to identify and classify new entries.

Maintain a centralized registry of known AI models by using a discovery workflow that fetches live provider data, updates the registry automatically, and supports cross-skill references to avoid hardcoded model IDs.

Yes, the model discovery tool works with local Ollama models alongside Anthropic, OpenAI, and Gemini, fetching their model names and classifying them into appropriate routing tiers.

AI workflows break when providers release new models because hardcoded model IDs become outdated, a problem solved by using a fetch/cache mechanism with a known-model registry that detects and classifies new models automatically.