search-specialist
Pass
Audited by Gen Agent Trust Hub on Sep 26, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill describes a research workflow that involves ingesting content from external websites (WebFetch Deep Dive), which is a surface for indirect prompt injection. While the instructions do not include specific boundary markers or sanitization steps for external data, the focus on multi-source verification and credibility assessment acts as a procedural mitigation against acting on untrusted instructions encountered during research.
Audit Metadata