vercel-react-best-practices
Pass
Audited by Gen Agent Trust Hub on Apr 29, 2026
Risk Level: SAFE
Full Analysis
- [METADATA_POISONING]: The skill's YAML frontmatter and internal documentation attribute authorship to 'Vercel' and 'Vercel Engineering', while the platform metadata identifies the publisher as 'udecode'. This discrepancy appears to be a matter of community attribution rather than a malicious attempt to deceive, as the content accurately reflects publicly available Vercel engineering guidelines.
- [INDIRECT_PROMPT_INJECTION]: The skill defines a large surface area for indirect prompt injection because its primary purpose is to have the agent process, refactor, and generate code based on untrusted user-provided React/Next.js projects. However, the guidelines provided are static, focus on performance best practices, and do not instruct the agent to ignore safety protocols.
- Ingestion points: User-provided React components and Next.js page files processed during refactoring or code review tasks.
- Boundary markers: None explicitly defined within the skill's instructions for separating guidelines from user code.
- Capability inventory: The skill instructs the agent to perform code generation, refactoring, and performance analysis.
- Sanitization: No specific sanitization or filtering instructions for handling user-provided code strings are provided.
- [EXTERNAL_DOWNLOADS]: The documentation references and recommends several well-known and standard Node.js packages including 'better-all', 'lru-cache', 'swr', 'zod', and '@vercel/analytics', as well as the 'svgo' CLI tool. All referenced resources are established industry standards for the described use cases.
Audit Metadata