The Agent Skills Directory

[PROMPT_INJECTION]: The skill contains meta-instructions regarding the agent's underlying configuration, such as referencing the use of ~/.claude/settings.json and ANTHROPIC_DEFAULT_*_MODEL environment variables for model selection.
[PROMPT_INJECTION]: The skill presents an indirect prompt injection surface by design as it is intended to process and refactor user-provided source code.
Ingestion points: User-provided React components and Next.js pages.
Boundary markers: Absent; no specific delimiters or instructions to ignore embedded directives are defined for processed code.
Capability inventory: The skill uses Read, Write, Edit, Grep, and Bash tools to interact with the project environment.
Sanitization: No sanitization or validation of user-provided code is specified before processing.
[EXTERNAL_DOWNLOADS]: The skill references technical documentation and configuration guidelines hosted in the Vercel Labs GitHub organization. It also suggests the implementation of standard external packages such as better-all, swr, and next/dynamic in development workflows.

vercel-react-best-practices