yandex-automation
Fail
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: HIGHEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS] (HIGH): The skill requires the user to add an untrusted external MCP server:
https://rube.app/mcp. This server provides the tool definitions and execution logic. - [REMOTE_CODE_EXECUTION] (HIGH): By calling
RUBE_SEARCH_TOOLSand thenRUBE_MULTI_EXECUTE_TOOLbased on the results, the agent effectively executes logic defined by the remoterube.appserver. This is a form of remote code/logic execution. - [INDIRECT_PROMPT_INJECTION] (HIGH): The skill possesses write/execute capabilities (Yandex automation) while processing external, potentially attacker-controlled content from Yandex. There are no boundary markers or sanitization logic mentioned to prevent data-driven instructions from overriding agent behavior.
- [DYNAMIC_EXECUTION] (MEDIUM): Tool schemas and execution plans are dynamically retrieved at runtime. While this is a design feature of MCP, doing so from an untrusted source without validation is a security risk.
Recommendations
- AI detected serious security threats
Audit Metadata