Maintains and updates model registries, pricing data, and capability maps for LLM backends and providers.
The Model Registry Maintainer skill provides a standardized framework for keeping AI model configurations accurate and current within the MassGen ecosystem. It guides developers through the complex process of updating model availability, release dates, pricing tiers, and context window limits across system-critical files like capabilities.py and token_manager.py. By integrating with the LiteLLM database and providing structured fallback patterns, this skill ensures that backend cost estimation, token counting, and feature validation stay synchronized with the fast-moving landscape of provider releases from OpenAI, Anthropic, Google, and xAI.
Key Features
01Precise management of context windows and input/output token limits
02Automated documentation and backend table regeneration scripts
03Automated pricing synchronization via LiteLLM database integration
04Standardized workflow for adding new LLM models and providers
05Capability mapping for vision, web search, and reasoning features
060 GitHub stars
Use Cases
01Configuring model-specific capability flags to enable advanced AI features
02Updating API pricing models to maintain accurate cost monitoring and budgeting
03Integrating a newly released model like GPT-5 or Claude 4 into an existing backend
