latex-paper-conversion

Warn

Audited by Gen Agent Trust Hub on Jun 22, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The skill instructions explicitly direct the agent to dynamically generate and execute a Python script (convert_format.py) to parse and transform LaTeX content. This pattern of script generation and execution on user-controlled data presents a risk of local code execution vulnerabilities.
  • [COMMAND_EXECUTION]: The skill requires the execution of system-level binary tools including pdflatex, bibtex, grep, and rg. Running compilers like pdflatex on untrusted input files can be dangerous, especially if features like shell-escape are enabled or if there are vulnerabilities in the TeX distribution.
  • [PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection.
  • Ingestion points: The skill reads untrusted .tex files provided by users (SKILL.md, Step 2).
  • Boundary markers: There are no instructions to use delimiters or warnings to ignore embedded instructions within the processed LaTeX files.
  • Capability inventory: The agent has the capability to write files, generate/execute Python scripts, and run shell commands (pdflatex, grep) based on the processed data.
  • Sanitization: No sanitization or validation of the LaTeX content is mentioned before it is processed by the generated Python script or the LaTeX compiler.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Jun 22, 2026, 09:57 AM
Security Audit — agent-trust-hub — latex-paper-conversion