ai-engineering

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly instructs use of OpenAI hosted web search and retrieval tools (see subskills/openai-web-search.md and subskills/retrieval.md) so the agent will fetch and consume public web pages/URLs as evidence that can influence answers and tool routing, exposing it to untrusted third‑party content.