shadcn-svelte

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

• Third-party content exposure detected (high risk: 0.85). The required workflow includes fetching live registry/docs from https://shadcn-svelte.com/registry/index.json and https://shadcn-svelte.com/llms.txt (and per-component JSON/docs) at runtime, which are outsider-authored public web content that can be ingested as free text into the agent’s LLM context.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly instructs fetching live registry and LLM-formatted docs at runtime (e.g., https://shadcn-svelte.com/registry/index.json, https://shadcn-svelte.com/llms.txt, and https://bits-ui.com/llms.txt), which are intended to be ingested/used by the agent and therefore can directly control prompts and behavior.