multi-search-engine
Pass
Audited by Gen Agent Trust Hub on Aug 19, 2026
Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from 17 external search engines, which could contain malicious instructions targeting the AI agent. The instructions lack explicit boundary markers or sanitization logic to isolate search results from the agent's core instructions. \n
- Ingestion points: Search results fetched via
web_fetchfrom URLs defined inconfig.jsonandSKILL.md. \n - Boundary markers: Absent. The skill instructions suggest organizing and summarizing results but do not provide delimiters or instructions to ignore embedded commands. \n
- Capability inventory: The
web_fetchtool is the primary capability used to query external domains. \n - Sanitization: Absent. The documentation does not define methods for filtering or escaping external content. \n- [EXTERNAL_DOWNLOADS]: The skill instructions specify fetching session cookies directly from the homepages of well-known search engines (such as Google, Baidu, and Bing) when search requests are rate-limited. These network operations target established, well-known services for the purpose of maintaining search session state.
Audit Metadata