The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill facilitates the installation and update of the vercel CLI package from the public NPM registry using pnpm. This is a standard procedure for utilizing Vercel's official developer tools.
[COMMAND_EXECUTION]: The skill's primary functionality is achieved by executing shell commands through the vercel binary. This includes operations for deployment, project linking, and resource management.
[DATA_EXPOSURE]: The skill manages sensitive data, including Vercel access tokens and environment variables. It incorporates a security best practice by explicitly advising the agent to use the VERCEL_TOKEN environment variable rather than the --token flag, which prevents secrets from appearing in process lists or command history.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface because it reads and processes data from external sources, such as build logs and project metadata (via vercel logs or vercel inspect). If these external sources contain malicious instructions, they could potentially influence the agent's logic.
Ingestion points: CLI output from vercel logs, vercel inspect, and vercel list (SKILL.md, references/commands.md).
Boundary markers: Not explicitly defined in the instructions.
Capability inventory: Shell command execution (vercel), local file writing (vercel env pull).
Sanitization: No specific filtering or escaping instructions are provided for the agent when processing CLI output.

vercel-cli