experiment
Pass
Audited by Gen Agent Trust Hub on Sep 6, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill is configured to connect to a remote Model Context Protocol (MCP) server at
https://mcp.trylexsis.com/mcp. This external dependency is required for the skill to access the Lexsis storefront tools and perform its primary functions. - [INDIRECT_PROMPT_INJECTION]: The skill instructions require the agent to ingest and analyze external data from the Lexsis platform to evaluate experiment performance and recommend actions.
- Ingestion points: Data retrieved from
lexsis_analytics.pageandlexsis_analytics.experimenttools, which read experiment status and metric movement from remote sources (File:SKILL.md). - Boundary markers: The instructions lack explicit delimiters or instructions to treat data from the analytics tools as untrusted, increasing the risk of the agent following instructions embedded in that data.
- Capability inventory: The skill includes significant mutation capabilities, including
lexsis_drafts.page_duplicate,lexsis_drafts.page_variation, andlexsis_drafts.experiment_create(File:SKILL.md). - Sanitization: There is no evidence of sanitization, validation, or schema enforcement on the data returned from the external analytics tools before it is processed for decision-making.
Audit Metadata