positioning
Pass
Audited by Gen Agent Trust Hub on Jun 17, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADS
Full Analysis
- [SAFE]: The skill implements a legitimate business workflow for strategic alignment. No patterns of obfuscation, credential theft, or unauthorized persistence were identified.\n- [EXTERNAL_DOWNLOADS]: The skill retrieves data from the web using well-known services such as Perplexity, Exa, and Firecrawl. These operations are limited to research purposes within the context of the positioning exercise and target official service endpoints.\n- [PROMPT_INJECTION]: The skill processes untrusted external content from competitor websites and user-provided documents. This creates a surface for indirect prompt injection; however, the structured nature of the analytical prompts (Phase 2 and Phase 3) focuses the agent on data extraction rather than execution, which mitigates the associated risk.
Audit Metadata