paper-writer

Pass

Audited by Gen Agent Trust Hub on Jun 22, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill implements a structured academic writing workflow that incorporates human-in-the-loop validation for each section. It includes clear instructions to the AI to avoid hallucinating research findings and to confirm recent publications using the 'WebSearch' tool.
  • [SAFE]: The 'pdf_to_images.py' script uses the well-known 'pymupdf' library to handle PDF processing tasks locally without external network dependencies within the script itself.
  • [SAFE]: Although the skill processes external PDF files, creating a surface for indirect prompt injection, it does so within a constrained and supervised context. Evidence Chain: Ingestion points include user-specified PDF files read through the 'Read' tool. Boundary markers are not specifically defined for PDF text ingestion. The capability inventory includes file editing and web access tools. No explicit sanitization of extracted PDF text is documented, but the interactive nature of the skill serves as a functional mitigation.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 22, 2026, 03:11 AM
Security Audit — agent-trust-hub — paper-writer