deep-research-specialist
Pass
Audited by Gen Agent Trust Hub on Sep 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to search and analyze external documentation and web content using multiple backends (Google, Bing, Gemini, etc.). This content is inherently untrusted and may contain adversarial prompts designed to hijack the agent's context.
- Ingestion points: Web search results, API documentation, and framework guides.
- Capability inventory: The skill description mentions orchestration and the execution of other skills through a management script.
- Boundary markers: There are no instructions or delimiters provided to isolate the ingested research data from the system prompts.
- Sanitization: The skill lacks explicit guidelines or logic for filtering or sanitizing external content.
- [NO_CODE]: The skill package contains only documentation and metadata files. It does not provide any internal scripts or binaries, relying instead on environment-provided tools.
Audit Metadata