The Agent Skills Directory

[SAFE]: The skill consists of markdown-based documentation and rule files that provide coding guidelines for React performance. No malicious logic, obfuscation, or safety bypasses were detected.
[DYNAMIC_EXECUTION]: The rule rendering-hydration-no-flicker.md suggests using dangerouslySetInnerHTML to inject an inline script that prevents visual flickering during hydration. This is a common React pattern for theme management and user preferences; the provided example uses a safe, hardcoded script template.
[INDIRECT_PROMPT_INJECTION]: The guidelines include patterns for reading from client-side data sources such as localStorage and searchParams. While these patterns define an attack surface for indirect prompt injection in the applications being developed, they are provided as legitimate performance optimizations and the skill includes guidance on authentication and validation (e.g., in rules/server-auth-actions.md).
[EXTERNAL_DOWNLOADS]: The documentation references reputable and well-known Node.js libraries such as swr, better-all, and lru-cache. It also points to official engineering blogs and documentation from Vercel and React, which are trusted sources.
[COMMAND_EXECUTION]: The README.md file defines standard development scripts (pnpm build, pnpm validate, pnpm extract-tests) for repository maintenance. These commands are typical for professional frontend development workflows and present no security risk to the agent or user environment.

vercel-react-best-practices