neverbounce-automation
Warn
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [Indirect Prompt Injection] (MEDIUM): The skill instructs the agent to fetch 'recommended execution plans' and 'tool schemas' from the Rube MCP server (
RUBE_SEARCH_TOOLS). This creates a vulnerability where the remote service could return malicious instructions that the agent might execute blindly. - [External Dependencies] (LOW): The skill requires the configuration of a third-party MCP endpoint (
https://rube.app/mcp). This service is not on the trusted providers list and acts as a gateway for all operations. - [Command Execution] (MEDIUM): The skill uses
RUBE_MULTI_EXECUTE_TOOLandRUBE_REMOTE_WORKBENCHto perform actions. These tools provide significant capabilities to modify data or perform remote operations based on schemas defined at runtime by the external server.
Audit Metadata