telnyx-ai-inference-curl

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

• Third-party content exposure detected (high risk: 1.00). The skill's SKILL.md explicitly includes an "Embed URL content" endpoint (POST /ai/embeddings/url) that crawls and loads website pages from a user-specified URL (and child pages) into a storage bucket, and those embedded pages can be searched/summarized (e.g., similarity-search, summarize, chat/embeddings), meaning untrusted public web content is ingested and can materially influence agent behavior.