The Agent Skills Directory

[SAFE]: No malicious prompt injection or instructions were detected. The skill is formatted as a structured set of guidelines for AI agents to improve application performance.
[SAFE]: External dependencies and references (e.g., swr, lru-cache, better-all) point to well-known, trusted technology services and open-source projects.
[SAFE]: The skill explicitly includes security-focused rules, such as server-auth-actions, which mandates authentication and authorization checks inside Server Actions to prevent unauthorized mutation access.
[SAFE]: Examples of sensitive operations, such as accessing localStorage or cookies, are used appropriately for client-side state persistence and performance optimizations without evidence of exfiltration or exposure.
[SAFE]: The use of dangerouslySetInnerHTML in the rendering-hydration-no-flicker rule is documented as a standard performance hack for theme synchronization and does not introduce vulnerabilities in the context provided.

vercel-react-best-practices