vercel-react-only-best-practices
Pass
Audited by Gen Agent Trust Hub on Oct 12, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONMETADATA_POISONING
Full Analysis
- [METADATA_POISONING]: The skill frontmatter identifies the author as 'vercel', which contradicts the system-provided author 'tianyili'. There is also a version mismatch between the metadata (1.1.0) and the version cited in the documentation (1.0.0).
- [INDIRECT_PROMPT_INJECTION]: The skill is intended to review and refactor user-provided React code, which is a known vector for instruction injection.
- Ingestion points: React components and data-fetching code supplied by the user for performance analysis.
- Boundary markers: The instructions do not define delimiters or directives for the agent to distinguish between its core task and instructions embedded in the user code.
- Capability inventory: The skill allows the agent to generate new React logic and perform automated refactoring.
- Sanitization: No input validation or sanitization is specified for the code context being processed.
Audit Metadata