vercel-react-best-practices
Pass
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: LOW
Full Analysis
- Indirect Prompt Injection (LOW): The skill triggers on tasks involving user-provided React/Next.js code. While this constitutes an ingestion surface where malicious instructions could be embedded in code comments or strings, the skill itself only provides static optimization rules and does not grant the agent expanded execution capabilities. Evidence: SKILL.md triggers and instructions.
- Unverifiable Dependencies & Remote Code Execution (LOW): Several rule files recommend the use of external npm libraries such as
better-all,lru-cache, andswr. These are reputable performance utilities within the React ecosystem. They are presented as architectural recommendations for the developer's project rather than automated runtime downloads for the agent itself. Evidence:rules/async-dependencies.md,rules/server-cache-lru.md. - Dynamic Execution (INFO): One guideline demonstrates a pattern for managing hydration states using
dangerouslySetInnerHTMLto inject an inline script. This is a standard industry practice for Next.js theme management and is presented here as a static, benign code example. Evidence:rules/rendering-hydration-no-flicker.md.
Audit Metadata