multi-agent-architecture

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

• Third-party content exposure detected (high risk: 0.90). The News Agent explicitly enables web_fetch and is described as "新闻抓取" with instructions to fetch news and include original links (see ~/.openclaw/agents/news/agent/config.json "tools": ["web_fetch"]), so the agent will ingest untrusted public web content that could carry indirect prompt injections.