plz-3d-explain
Pass
Audited by Gen Agent Trust Hub on Oct 30, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructions require the agent to run a local HTTP server using
python3 -m http.serverfor verifying generated HTML pages. The server is configured to bind strictly to127.0.0.1to ensure it is not accessible from the network. - [DYNAMIC_EXECUTION]: The agent is directed to use browser automation tools like Playwright to interact with the generated pages. This allows for automated verification of the 3D scene layout, such as checking for overlapping labels or rendering issues.
- [EXTERNAL_DOWNLOADS]: The generated output and templates load the Three.js library and various typography assets from reputable and well-known CDNs, including JSDelivr and Google Fonts. These references are documented and essential for the functionality of the 3D models.
- [INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted technical data from source repositories or web search results to create explanations. To mitigate risks, the instructions mandate that the agent performs fact-checking, cites sources with verification dates, and clearly identifies any information that remains unverified.
Audit Metadata