naval-perspective
Pass
Audited by Gen Agent Trust Hub on Apr 9, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The skill includes directives to maintain role-play consistency by suppressing typical meta-analysis and repeated disclaimers. These instructions are focused on persona fidelity and do not attempt to bypass core safety filters or extract system prompts.
- [INDIRECT_PROMPT_INJECTION]: The agent is instructed to use a web search tool to research specific topics before responding. This processing of untrusted third-party data represents an indirect injection surface.
- Ingestion points: Results from the WebSearch tool (e.g., product feedback, market trends) in SKILL.md.
- Boundary markers: No specific delimiters are used to wrap external content.
- Capability inventory: The agent uses search tools and provides analytical responses based on retrieved data.
- Sanitization: No explicit content filtering or validation of search results is defined in the instructions.
Audit Metadata