vercel-react-best-practices
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [External Downloads] (LOW): The documentation recommends using 'pnpm install' for development and references several standard npm packages such as swr, lru-cache, and zod. Given that the skill is authored by Vercel Engineering, which is a trusted organization, these download references are downgraded to LOW per the Trusted Source rule.
- [Prompt Injection] (LOW): The skill is designed to process user-provided code for refactoring, creating a surface for Indirect Prompt Injection. 1. Ingestion points: Untrusted React or Next.js source code. 2. Boundary markers: There are no specific instructions or delimiters provided to prevent the agent from following malicious instructions embedded in the code it refactors. 3. Capability inventory: An agent using this skill likely possesses file-write and code-execution permissions. 4. Sanitization: No guidance is provided for sanitizing the processed code for hidden instructions.
- [Dynamic Execution] (LOW): The 'rendering-hydration-no-flicker.md' rule provides an example of using dangerouslySetInnerHTML to inject a synchronous script. While this is a common React pattern for theme hydration, it constitutes a low-risk form of dynamic script generation from a static template.
- [Data Exposure & Exfiltration] (SAFE): No hardcoded secrets, sensitive file paths, or malicious exfiltration patterns were detected. Examples of localStorage and fetch usage are consistent with standard web development practices.
Audit Metadata