vercel-composition-patterns
Warn
Audited by Gen Agent Trust Hub on Sep 26, 2026
Risk Level: MEDIUMMETADATA_POISONINGEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTION
Full Analysis
- [METADATA_POISONING]: The skill's metadata claims the author is "vercel", despite being authored by "pymodel". This is a deceptive practice that misrepresents the skill's provenance and may lead users to grant it undue trust.
- [EXTERNAL_DOWNLOADS]: The
SKILL.mdfile contains instructions to executenpx -y @pymodel/react-frontend-skills, which triggers a download of a third-party package from the npm registry. - [REMOTE_CODE_EXECUTION]: By recommending the use of
npx -y, the skill encourages the execution of remote code from a non-trusted source without human review or confirmation prompts. The risk is compounded by the deceptive metadata which attempts to brand the tool as an official Vercel resource.
Audit Metadata