deep-research-agent

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

• Third-party content exposure detected (high risk: 0.90). The skill's required workflow (SKILL.md Step 3 "执行研究（增量）" and the examples/research-workflow-example.md) explicitly instructs the agent to search for public web sources, save source entries with URLs into materials/ (e.g., file_write including "URL: https://...") and then use those third-party findings to update report.md and drive subsequent actions, which means it fetches/ingests open public content that can materially influence tool use and decisions.