The Agent Skills Directory

[EXTERNAL_DOWNLOADS] (SAFE): The skill references standard, trusted commands such as npx shadcn@latest. These are directed to well-known official registries and the shadcn project ecosystem.
[COMMAND_EXECUTION] (SAFE): Includes a shell script (scripts/verify-setup.sh) which performs read-only checks (using grep, find, and ls) on the local project structure to ensure proper configuration. It does not perform any destructive or unauthorized operations.
[PROMPT_INJECTION] (SAFE): The SKILL.md and README.md contain standard instructional guidelines for a technical assistant. There are no attempts to bypass safety filters or override system instructions.
[REMOTE_CODE_EXECUTION] (SAFE): While the skill assists with npx commands, these are restricted to the official shadcn CLI tools, which is the standard and documented method for using this library.
[DATA_EXFILTRATION] (SAFE): No network operations to unknown or suspicious domains were found. All tools used (like web_fetch) are restricted to standard documentation and registry exploration.

shadcn-ui