The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill references several well-known and trusted external libraries and resources, including:
swr and better-all (associated with Vercel and its engineering team).
lru-cache (a widely used utility from a known developer).
Popular UI and utility libraries such as lucide-react, @mui/material, and lodash.
Documentation links pointing to official domains like react.dev, nextjs.org, and vercel.com.
[COMMAND_EXECUTION]: The README.md provides standard project maintenance instructions including pnpm install, pnpm build, pnpm validate, and pnpm extract-tests. Additionally, the rendering rules suggest using npx svgo for SVG optimization.
[PROMPT_INJECTION]: The skill includes instructions in AGENTS.md and SKILL.md designed to guide agent behavior when performing code refactoring. These are domain-specific performance rules rather than attempts to bypass security filters.
[PROMPT_INJECTION]: Indirect Prompt Injection Surface:
Ingestion points: The agent is instructed to use these guidelines when reviewing or refactoring user-provided React/Next.js codebases (e.g., AGENTS.md abstract).
Boundary markers: No specific delimiters or instructions for isolating user code are provided in the skill content.
Capability inventory: The agent is tasked with generating and modifying source code based on these rules.
Sanitization: The skill does not provide methods for sanitizing the user code it processes.

react-best-practices