DollhouseMCP/mcp-server

DollhouseMCP aims to provide a centralized hub for AI personas and skills using the Model Context Protocol (MCP) standard recently popularized by Anthropic. While it addresses a real need—standardizing how LLMs adopt specific roles and toolsets across different clients (like Cursor, Claude Desktop, or custom IDEs)—it lacks a significant technical or community moat. With only 29 stars and 10 forks over nearly 300 days, the project has not achieved the network effects required to make its 'persona collection' a true asset. The functionality is largely a wrapper around standard MCP patterns. This space is under extreme threat from frontier labs: Anthropic's 'Projects' and OpenAI's 'GPTs' already provide native persona and tool management. Furthermore, AI-native IDEs like Cursor and Windsurf are increasingly building their own internal registries for system instructions, which could render standalone persona servers redundant. The project's value is currently as a reference implementation for MCP enthusiasts, but it faces rapid displacement as the MCP ecosystem matures and more robust, official registries emerge.