skills/liangmiqwq/skills/antfu-design/Gen Agent Trust Hub

antfu-design

Pass

Audited by Gen Agent Trust Hub on Oct 1, 2026

Risk Level: SAFEPERSISTENCECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDATA_EXFILTRATIONMETADATA_POISONING
Full Analysis
  • [PERSISTENCE]: The skill instructs the agent to modify the project's AGENTS.md file by appending a rule regarding component reuse. This ensures that the design system guidelines persist across subsequent agent interactions within the project environment.
  • [COMMAND_EXECUTION]: Documentation references the use of a CLI contrast scanner tool (tsx .../a11y/cli.ts <url>). While this involves command execution on potentially external URLs, it is framed as a standard development utility for accessibility testing.
  • [EXTERNAL_DOWNLOADS]: The skill recommends installing several standard development packages via pnpm, including @antfu/design, unocss, storybook, and related Vue/Vite plugins. All identified packages are standard in the frontend development ecosystem.
  • [DATA_EXFILTRATION]: The DisplayIconifyRemoteIcon component performs network requests to the Iconify API to fetch icon data dynamically. This is a standard and documented feature of the design system for rendering remote icons.
  • [METADATA_POISONING]: The skill metadata identifies 'antfu' as the author, whereas the system identifies the author as 'liangmiqwq'. This discrepancy appears to be an attribution to the library creator rather than a malicious attempt to deceive.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 1, 2026, 02:29 PM