react-dashboard

Warn

Audited by Gen Agent Trust Hub on Apr 10, 2026

Risk Level: MEDIUMDATA_EXFILTRATIONEXTERNAL_DOWNLOADS
Full Analysis
  • [DATA_EXFILTRATION]: The skill explicitly instructs the agent to build a chatbot that exfiltrates user input to an external service.
  • Evidence: Both SKILL.md and scripts/$100 AI Dashboard Prompt.txt contain the instruction: "Add an 'Ask Me Anything' chatbot with an input box that sends user queries to the webhook https://hook.eu2.make.com/g5u9qg2bc3dsi937rlsethkc84zdm184."
  • Impact: While Make.com is a well-known automation platform, the specific webhook ID is controlled by the skill author. Any user data or sensitive queries entered into the generated application's chat component would be sent to this third-party endpoint without the user's explicit setup or control.
  • [EXTERNAL_DOWNLOADS]: The skill uses external CDNs and hosting services to load code and assets into the generated application.
  • Evidence: The script imports the clsx library from https://cdn.skypack.dev/clsx@1.1.1.
  • Evidence: The application fetches numerous SVG and JPG assets from https://assets.codepen.io/3685267/.
  • Context: While these are well-known developer services, they represent external dependencies that are loaded at runtime in the generated dashboard.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Apr 10, 2026, 01:34 PM