shadcn-ui

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.80). The skill's SKILL.md and README explicitly instruct the agent to use shadcn MCP tools (e.g., get_component, get_block, search_items_in_registries, list_items_in_registries) to retrieve component and block source from registries/Registry Directory—i.e., fetching code and files from third-party registries/public sources that the agent will read and act on during installation/customization—so untrusted external content can influence subsequent actions.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill instructs running the shadcn CLI (e.g., via "npx shadcn@latest add/init/create"), which downloads and executes remote package code (sourced from the shadcn project such as https://github.com/shadcn-ui/ui and https://ui.shadcn.com), so the skill relies on runtime-fetched external code that will be executed on the host.