physics-writing-revtex

Pass

Audited by Gen Agent Trust Hub on Sep 7, 2026

Risk Level: SAFENO_CODE
Full Analysis
  • [SAFE]: The skill consists entirely of natural language instructions for a LaTeX-based workflow. No executable code, scripts, or automated network commands are included.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided paper content, which presents a surface for indirect prompt injection if the content contains malicious LaTeX commands. However, the skill does not specify unsafe compilation flags (like --shell-escape), and this is an inherent risk of the LaTeX format rather than a malicious pattern in the skill itself.
  • Ingestion points: Processes 'paper content' provided by the user (SKILL.md).
  • Boundary markers: None specified.
  • Capability inventory: Instructs the agent to compile files using pdfLaTeX or LuaLaTeX and prepare file packages (SKILL.md).
  • Sanitization: None specified.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 7, 2026, 02:59 AM
Security Audit — agent-trust-hub — physics-writing-revtex