rlm

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

• Third-party content exposure detected (high risk: 1.00). The skill explicitly ingests arbitrary public web content — e.g., "rlm ask https://x.com" (auto-converts URLs to markdown) and the Exa web search API (exa.search/web and exa.find_similar with include_text/highlights) — so the agent will fetch and read untrusted third-party pages as part of its workflow.