ocr-and-documents
Pass
Audited by Gen Agent Trust Hub on Sep 5, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill implements document parsing and OCR functionalities using standard, well-known Python libraries (pymupdf, pymupdf4llm, marker-pdf).
- [EXTERNAL_DOWNLOADS]: The skill mentions marker-pdf downloads models (~2.5GB) from Hugging Face's official repository at
~/.cache/huggingface/on first use. This is standard behavior for machine learning models and follows established security practices. - [COMMAND_EXECUTION]: The skill uses
execute_codeand shell scripts to perform legitimate document operations like merging, splitting, and extracting text/tables. These operations are scoped to the user-provided files. - [INDIRECT_PROMPT_INJECTION]: As a document extraction skill, it processes untrusted external files (PDFs, scans). While the extraction process itself is safe, the extracted text is intended for agent context. The skill correctly recommends standard parsing libraries over manual regex for complex formats like Word or PowerPoint, reducing the risk of schema confusion.
Audit Metadata