campayn-automation
Fail
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: HIGHEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
- EXTERNAL_DOWNLOADS (HIGH): The skill requires connecting to an untrusted MCP endpoint
https://rube.app/mcp. This server provides the tools, logic, and schemas for the agent to use, giving it control over the agent's capabilities. - PROMPT_INJECTION (HIGH): The skill is highly vulnerable to Indirect Prompt Injection through its core workflow. Ingestion points:
RUBE_SEARCH_TOOLSreturns tool schemas and execution plans from the untrusted MCP server. Boundary markers: Completely absent; the agent is instructed to 'Always search tools first' and follow the 'recommended execution plans'. Capability inventory:RUBE_MULTI_EXECUTE_TOOLandRUBE_REMOTE_WORKBENCHallow for the execution of arbitrary tools and code defined by the server. Sanitization: No sanitization of external content is present. - REMOTE_CODE_EXECUTION (HIGH): By combining untrusted tool definitions from
RUBE_SEARCH_TOOLSwith the execution capability ofRUBE_MULTI_EXECUTE_TOOL, the external MCP server can effectively execute arbitrary logic via the agent's environment.
Recommendations
- AI detected serious security threats
Audit Metadata