shadcn-ui

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

• Third-party content exposure detected (high risk: 0.90). The skill's SKILL.md and vendor/shadcn/SKILL.md / cli.md explicitly instruct the assistant to run commands like "npx shadcn@latest docs" and use MCP/view to fetch component docs, example URLs, registry items and raw files (e.g., ui.shadcn.com, GitHub raw, community registries), which are untrusted public third‑party sources whose content the agent is expected to read and that directly influence install/update/apply decisions and CLI actions.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly instructs the agent at runtime to fetch component documentation/example URLs (e.g. https://raw.githubusercontent.com/.../examples/button-example.tsx) and use that fetched content as authoritative guidance for generating code, so remote raw GitHub content can be fetched and injected into the agent's context and directly control prompts.