edu-solid-geometry-doubao

Pass

Audited by Gen Agent Trust Hub on Sep 29, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied geometry problem text and images to generate educational content. This creates an attack surface where untrusted data enters the agent context. However, the skill utilizes a structured JSON schema to manage data interpolation and performs validation through scripts/validate.py to ensure consistency.\n- [COMMAND_EXECUTION]: The skill contains a test utility (scripts/selftest.py) that employs subprocess.run to execute internal scripts for regression testing. This functionality is scoped to the skill's own directory and is used to verify the solver and builder components.\n- [EXTERNAL_DOWNLOADS]: The generated HTML template fetches essential libraries like Three.js and MathJax from reputable Content Delivery Networks, including Cloudflare, Google, UNPKG, and jsDelivr. These references are standard for web-based 3D visualization and do not involve untrusted remote code execution.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 29, 2026, 09:28 AM