Arcana serves as your codebase's institutional memory, an open-source MCP server designed to embed your team's collective knowledge—including decisions, incidents, conventions, and ownership—into the code review process. It functions as an AI reviewer that retains historical context, leveraging Tree-sitter for deep codebase understanding, dependency graphs, and symbol relationships. By recalling past events and insights, Arcana enables more informed, contextual code reviews and provides powerful querying capabilities for architects and developers.
使用案例
01Perform impact analysis on code changes to identify affected areas.
02Conduct contextual code reviews enhanced with historical decisions and incident data.
03Query the codebase in natural language for architectural decisions, conventions, and past incidents.
