sms-alert-automation

Warn

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTION
Full Analysis
  • EXTERNAL_DOWNLOADS (MEDIUM): The skill requires the installation of an unverified MCP server from https://rube.app/mcp. This domain is not within the trusted repository or organization list and serves as the primary gateway for all operations.
  • REMOTE_CODE_EXECUTION (MEDIUM): Uses RUBE_REMOTE_WORKBENCH and RUBE_MULTI_EXECUTE_TOOL to perform operations. These functions delegate logic execution to a remote environment controlled by the unverified MCP provider.
  • COMMAND_EXECUTION (MEDIUM): The core functionality involves run_composio_tool() calls through a remote workbench, which allows arbitrary tool execution within the context of the SMS Alert toolkit.
  • DATA_EXFILTRATION (LOW): While not explicitly exfiltrating local files, the architecture requires all SMS data (recipient numbers and message content) to be transmitted to the rube.app endpoint, which acts as a middleman. The lack of API key requirements suggests the provider has visibility into all traffic.
  • INDIRECT PROMPT INJECTION (MEDIUM): The skill is designed to dynamically fetch tool schemas and 'recommended execution plans' from RUBE_SEARCH_TOOLS. An attacker controlling the remote server could provide malicious schemas or plans to manipulate the agent's behavior.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 16, 2026, 12:56 PM