The Agent Skills Directory

[PROMPT_INJECTION]: The skill identifies a surface for indirect prompt injection by processing untrusted external data from user-uploaded documents (PDF, TXT, Word). * Ingestion points: The skill reads user-provided files from the 'uploads/' directory. * Boundary markers: No specific delimiters or system instructions are provided to distinguish between document content and agent instructions. * Capability inventory: The skill utilizes file system access to read from 'uploads/' and write analysis results to the 'analysis/' and 'data/' directories. * Sanitization: There is no mention of sanitizing or escaping the content extracted from documents before it is processed by the AI model.

document-analysis