The Agent Skills Directory

[SAFE]: The skill is designed to analyze user reading data (highlights and documents) using official Readwise MCP tools or the readwise CLI. This data access is core to the skill's purpose and is performed within the user's existing security context. While the skill ingests external content, it lacks the capabilities (such as arbitrary network access or file system writes) required to exploit potential indirect prompt injections. All operations are confined to read-only data retrieval and natural language analysis.

surprise-me