plain-automation

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.80). The skill instructs the agent to contact the external Rube MCP server (https://rube.app/mcp) and to call RUBE_SEARCH_TOOLS / RUBE_GET_TOOL_SCHEMAS which return external tool schemas and recommended execution plans that the agent must read and interpret, exposing it to untrusted third-party content that could carry indirect prompt injection.