research-methodology
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from user research questions and external sources via web tools, creating an indirect prompt injection surface.
- Ingestion points: User-supplied research problems and external content fetched from the web using the WebFetch tool.
- Boundary markers: The instructions lack specific delimiters or warnings to ignore commands embedded in the research data.
- Capability inventory: The skill utilizes WebSearch, WebFetch, Read, Grep, and Glob tools for data analysis.
- Sanitization: No validation or sanitization of input data is specified in the protocol.
Audit Metadata