multi-source-search

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

• Third-party content exposure detected (medium risk: 0.30). In scripts/search_client.py, outsider-supplied --query/--system text is sent to AIsa’s Perplexity “sonar/sonar-pro/sonar-reasoning-pro/sonar-deep-research” endpoints (and Tavily search/crawl/extract), which perform web retrieval and ingest outsider-authored page/comment content into the generated answer at runtime.