chinese-stem-terminology

Pass

Audited by Gen Agent Trust Hub on Sep 20, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill operates entirely on local markdown reference files (references/mathematics.md, references/physics.md, etc.). It contains no code, executable scripts, or network operations.
  • [SAFE]: Instructions specifically warn the agent against making authoritative claims (e.g., claiming a term is the "only correct" standard) and mandate the protection of technical entities like URLs, file paths, and code blocks.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided STEM text for translation or proofreading.
  • Ingestion points: The skill reads user input prompts as defined in evals.json and trigger-evals.json.
  • Boundary markers: The SKILL.md defines a "Protection" section to explicitly ignore internal characters in formulas, code, and URLs, which serves as a basic safety boundary.
  • Capability inventory: The skill has no capabilities for file writing, network access, or subprocess execution; it is a text-transformation prompt only.
  • Sanitization: No explicit sanitization of input is performed, but the lack of dangerous tools minimizes the risk of exploitation.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 20, 2026, 02:55 AM
Security Audit — agent-trust-hub — chinese-stem-terminology