The Agent Skills Directory

Command Execution: The skill is designed to run shell commands such as npx, npm, and yarn to install dependencies and execute codemods. These are standard operations for JavaScript project maintenance but involve executing code on the local system.
External Data Ingestion and Indirect Prompt Injection Surface: The instructions direct the agent to fetch documentation from nextjs.org using WebFetch to determine the appropriate upgrade path. This creates a surface for indirect prompt injection, where content from the external website could theoretically influence the parameters passed to shell commands.
Ingestion points: Documentation is fetched from nextjs.org URLs defined in SKILL.md.
Boundary markers: No specific delimiters or instructions to ignore embedded commands in the fetched content are defined.
Capability inventory: The skill can execute shell commands (npx, npm install, npm run) as defined in SKILL.md.
Sanitization: No explicit sanitization or validation of the fetched documentation is specified before it is used to determine command arguments.
Remote Code Execution: The skill uses npx @next/codemod@latest, which downloads and executes the latest version of the official Next.js codemod utility from the npm registry. While this is the intended functionality for the upgrade process and uses official tooling, it represents the execution of remote code.

next-upgrade