qoder-wiki

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

• Third-party content exposure detected (high risk: 1.00). The skill explicitly tells the agent to "如文档信息不足，再考虑联网搜索补充" in SKILL.md and the docs (e.g., docs/拓展能力/MCP.md and docs/拓展能力/自定义智能体.md) expose WebFetch/WebSearch and MCP-based webpage fetching of arbitrary URLs, so the agent will fetch and interpret untrusted public web content (web pages/search results) that can materially influence tool use and decisions.