Pass
Audited by Gen Agent Trust Hub on Apr 10, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The read_paper function in tools.py uses the httpx library to download PDF files from URLs provided by the user.
- [COMMAND_EXECUTION]: The SKILL.md guide provides examples and instructions for using system-level PDF tools such as qpdf, pdftotext, and pdftk, which involves executing shell commands.
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it extracts text from untrusted PDF files that may contain malicious instructions targeting the AI agent.
- Ingestion points: Text extraction from both local and downloaded PDF files in tools.py using pdfplumber and PyPDF2.
- Boundary markers: The skill does not implement specific delimiters or instructions to help the agent distinguish between its instructions and the extracted document content.
- Capability inventory: The skill possesses capabilities for network downloads, file system writes (saving PDFs and images), and shell command execution.
- Sanitization: Extracted text is returned directly to the agent context without any filtering or validation.
Audit Metadata