quiz
Pass
Audited by Gen Agent Trust Hub on Nov 6, 2026
Risk Level: SAFEDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [DATA_EXFILTRATION]: The skill connects to an external Model Context Protocol (MCP) server at
https://mcp.trylexsis.com/mcpto manage quiz storefront resources and capture policies. This involves network communication with a non-whitelisted domain. - [INDIRECT_PROMPT_INJECTION]: The skill handles merchant-authored quiz definitions and shopper-provided answers which are interpolated into storefront content and navigation URLs, creating a surface for indirect prompt injection.
- Ingestion points: Quiz definition source code and shopper answers collected during the interaction.
- Boundary markers: The skill uses structured JSON for definitions but lacks explicit delimiters or warnings to the AI to ignore embedded instructions within these inputs.
- Capability inventory: Creating and updating unpublished draft pages, activating data capture policies, and publishing storefront pages via the Lexsis MCP tools.
- Sanitization: Documentation specifies that shopper-provided placeholders are URL-encoded and that personal or sensitive data is redacted from the MCP tool output.
Audit Metadata