skills/diegosouzapw/omniroute/cli-mcp/Gen Agent Trust Hub

cli-mcp

Warn

Audited by Gen Agent Trust Hub on Aug 24, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructs the user to perform a global installation of the omniroute package from the npm registry without specifying a version. This package is not provided by a recognized trusted vendor or well-known service.
  • [COMMAND_EXECUTION]: The skill relies on running the omniroute CLI to interact with the Model Context Protocol (MCP) server. It exposes commands to list tools, restart the server, and invoke specific tools.
  • [INDIRECT_PROMPT_INJECTION]: The skill provides an interface to invoke tools with JSON arguments, which constitutes an indirect injection surface. * Ingestion points: The argsJson and --args parameters of the mcp call subcommand found in SKILL.md. * Boundary markers: None provided in the command structure or instructions. * Capability inventory: Ability to invoke arbitrary tools and potentially influence system state through the omniroute CLI. * Sanitization: There is no mention of sanitization or validation for the JSON arguments passed to the tools.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Aug 24, 2026, 05:24 PM
Security Audit — agent-trust-hub — cli-mcp