ocr-scanned-pdf
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute the
ocrmypdfandpdftotextbinaries. These are standard open-source tools for PDF processing and are used here for their intended purpose of checking for text layers and performing OCR. - [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted PDF files to perform OCR. While these files could theoretically contain malicious content designed to influence the agent's subsequent processing of the extracted text, the skill does not contain any specific patterns that increase this risk beyond standard file processing.
- Ingestion points: Processes user-provided PDF file paths (SKILL.md).
- Boundary markers: None explicitly defined within the skill instructions.
- Capability inventory: Executes external command-line tools (ocrmypdf, pdftotext) and writes to the local file system.
- Sanitization: Not explicitly defined; relies on the underlying tools' handling of PDF content.
Audit Metadata