概要
This skill provides specialized guidance for maintaining the core model registry of the MassGen system, ensuring the multi-agent framework remains current with the latest frontier model releases. It streamlines the process of updating model capabilities in the backend registry and managing complex pricing or context window data in the token manager. By leveraging both hardcoded overrides and the LiteLLM database, this skill helps developers keep the system's cost estimation and model selection features accurate as providers like OpenAI, Anthropic, and Google release new iterations.