AIButton

PassAudited by VirusTotal on Apr 2, 2026.

Findings (1)

The AIButton skill bundle is classified as suspicious because SKILL.md contains instructions for the AI agent to execute high-risk shell commands (e.g., rm -rf, zip) and references specific local filesystem paths (e.g., /Users/lippsli/Desktop/AI控件文档/). While these actions are plausibly intended for packaging the generated UI components, the use of shell execution and hardcoded paths poses a risk in an agentic environment. Additionally, the technical specifications for the generated index.html include an insecure postMessage implementation using a wildcard origin ('*'), which is a security vulnerability.