The Agent Skills Directory

[Metadata Analysis] (SAFE): The skill description claims to synchronize context and run at launch, which implies a persistence or automation role, but the provided commands are limited to local document conversion.
[Indirect Prompt Injection] (LOW): The skill processes external data via .docx files. While it lacks explicit sanitization or boundary markers for the content of these files, the capabilities are limited to reading the converted output to the console, posing minimal risk.
[Command Execution] (SAFE): The commands pandoc and cat are used for their intended purposes of document processing and file reading. No dangerous flags or shell escapes were detected.

context-loader