shadcn-svelte

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

• Third-party content exposure detected (high risk: 0.80). The skill's CLI explicitly fetches remote components and registries (see references/cli.md: the "add" command accepts component names or a URL and supports a configurable "registry" in components.json and proxy/outgoing requests), meaning it ingests content from open/public third‑party sources which the agent would read/interpret to add components.