market-landing
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests content from external URLs for analysis, creating a surface for potential indirect prompt injection. While malicious content on a target page could attempt to influence the agent, the skill provides comprehensive frameworks and ethical constraints that focus the agent on objective CRO metrics.\n
- Ingestion points: Landing page content accessed through user-provided URLs in the trigger commands (SKILL.md).\n
- Boundary markers: There are no explicit markers used to distinguish external data from instructions.\n
- Capability inventory: Capability to write output files (LANDING-CRO.md) and potentially interact with browser tools for screenshots.\n
- Sanitization: No specific sanitization or instruction-filtering process is defined for the external data.
Audit Metadata