Implements standardized patterns for exposing read-only data as Model Context Protocol (MCP) resources using URI-based identification.
This skill provides a comprehensive framework for designing and implementing MCP resources, enabling Claude to access read-only context such as documentation, database schemas, and logs. It guides developers through defining custom URI schemes, implementing efficient list_resources and read_resource handlers, and ensuring security via path validation and MIME type detection. By strictly separating read-only context from side-effect actions, this skill ensures that AI agents interact with system data in a predictable, secure, and highly structured manner.
主な機能
01Standardized URI scheme design for files, databases, logs, and custom schemas
02Robust security validation to prevent path traversal and unauthorized access
03Automated MIME type detection and support for binary resource encoding
04Asynchronous implementation patterns for list_resources and read_resource handlers
05Dynamic resource generation for time-sensitive data like application logs
060 GitHub stars
ユースケース
01Providing read-only access to database schemas for improved AI-driven SQL generation
02Provisioning application logs for automated error analysis and debugging sessions
03Exposing project documentation and API specifications to Claude as searchable context
