ipdata-co-automation

Fail

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: HIGHREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • Indirect Prompt Injection (HIGH): The skill creates a significant attack surface by consuming external data and schemas from the Ipdata.co API and the Rube MCP server. These outputs are used to drive the RUBE_MULTI_EXECUTE_TOOL and RUBE_REMOTE_WORKBENCH functions.
  • Ingestion Points: Data enters the agent context via RUBE_SEARCH_TOOLS (tool schemas) and the output of executed Ipdata tools.
  • Capability Inventory: The skill has broad tool execution capabilities (RUBE_MULTI_EXECUTE_TOOL) and remote workbench access (RUBE_REMOTE_WORKBENCH).
  • Boundary Markers: None identified in the provided markdown; instructions do not explicitly warn the agent to ignore instructions embedded in the IP data or tool schemas.
  • Sanitization: No evidence of sanitization or schema validation is present in the skill definition.
  • Remote Code Execution / External Downloads (HIGH): The setup instructions require the user to add https://rube.app/mcp as an MCP server. This endpoint is not a trusted source according to the security framework. This server dynamically provides tool definitions and execution logic to the agent, effectively serving as a remote execution vector.
  • Command Execution (MEDIUM): The use of RUBE_REMOTE_WORKBENCH with run_composio_tool() allows for complex, potentially dangerous operations that are abstracted behind the MCP interface, reducing transparency into what commands are actually being executed.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Feb 17, 2026, 12:29 AM