The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill includes standard commands to install and upgrade framework dependencies using official package managers (npm, npx).
Evidence: npm install next@latest react@latest react-dom@latest and npm install babel-plugin-react-compiler@latest.
Evidence: npx @next/codemod@canary upgrade latest and npx create-next-app@latest.
Context: These are official tools and packages for the Next.js ecosystem.
[COMMAND_EXECUTION]: The skill provides instructions for standard development workflows including building, testing, and running the development server.
Evidence: npm run dev, npm run build, and npm test.
Context: These are routine commands for JavaScript/TypeScript project management.
[DATA_EXPOSURE]: The skill includes code snippets demonstrating how to handle cookies and headers using new asynchronous APIs.
Evidence: Examples showing await cookies() and await headers().
Context: These snippets follow the evolution of the framework's API and do not expose sensitive hardcoded credentials or private data.
[INDIRECT_PROMPT_INJECTION]: The skill describes an integration with the Model Context Protocol (MCP) for AI-assisted debugging, creating a surface where an agent reads project logs and code.
Ingestion points: Agent reading server logs and route configuration via DevTools MCP (SKILL.md).
Boundary markers: None explicitly mentioned in the guide content.
Capability inventory: Shell execution for package management and build processes (SKILL.md).
Sanitization: Not applicable as the content is educational documentation.
Context: As a documentation skill, the exposure is limited to standard development context and poses no unique threat.

nextjs-16-complete-guide