sci-paper-reviewer

Pass

Audited by Gen Agent Trust Hub on May 20, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill includes a Python script scripts/enhanced_document_parser.py intended to be run locally to extract text from manuscripts. The script performs legitimate file reading and parsing operations using standard libraries and well-known third-party packages.
  • [EXTERNAL_DOWNLOADS]: The skill documents dependencies on several widely used document parsing libraries available via standard package registries (PyPI), including PyPDF2, pdfplumber, pymupdf, pypdf, and python-docx.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes text extracted from external, untrusted sources (user-provided manuscripts).
  • Ingestion points: User-uploaded manuscripts are processed by scripts/enhanced_document_parser.py and the extracted text is then analyzed by the agent.
  • Boundary markers: The skill lacks explicit markers or delimiters (e.g., XML tags or clear separators) to encapsulate the extracted text, increasing the risk that instructions embedded in a manuscript could be interpreted as agent commands.
  • Capability inventory: The agent has the capability to read local files and generate structured reports based on the manuscript content.
  • Sanitization: No sanitization or validation of the extracted text is performed before it is provided to the agent for review.
Audit Metadata
Risk Level
SAFE
Analyzed
May 20, 2026, 01:51 AM
Security Audit — agent-trust-hub — sci-paper-reviewer