deepsearch

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

• Third-party content exposure detected (high risk: 0.90). The SKILL.md explicitly depends on a web-search skill that runs python scripts/web_search.py "<query>" to fetch and append real web search results into the findings array, and those untrusted public web contents are read by the LLM to decide nextSearchTopic and drive further searches and analysis.