arxiv-translator

Pass

Audited by Gen Agent Trust Hub on Aug 8, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill downloads LaTeX source files and reference PDFs from arxiv.org, which is a recognized and trusted service for scientific publications.
  • [COMMAND_EXECUTION]: Standard utilities including curl, tar, xelatex, and bibtex are invoked to manage the automated retrieval and compilation of documents.
  • [REMOTE_CODE_EXECUTION]: A locally provided Python script (scripts/split_tex.py) is executed to handle the modularization of LaTeX files; this script performs standard text manipulation without any network exfiltration or obfuscation.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection via the external LaTeX source it processes. Ingestion points: LaTeX content is read into the agent context through the file-splitting script. Boundary markers: No explicit delimiters or ignore-instructions are defined. Capability inventory: The skill can execute shell commands, write files, and run Python code. Sanitization: The source text is translated and compiled without prior sanitization or filtering.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 8, 2026, 03:03 AM
Security Audit — agent-trust-hub — arxiv-translator