skills/robsonrung/rar-skills/cmux-cli/Gen Agent Trust Hub

cmux-cli

Pass

Audited by Gen Agent Trust Hub on Sep 8, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses the cmux command-line utility to manage terminal workspaces, panels, and input. These commands are restricted to the local environment and are consistent with the skill's documented purpose of terminal management.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests JSON output from tool commands (e.g., cmux identify --json, cmux list-panels --json) to resolve workspace and surface IDs. While this constitutes an attack surface for data-driven instructions, the risk is mitigated by explicit instructions to verify state changes with independent queries and the focus on observable behavior.
  • [EXTERNAL_DOWNLOADS]: The skill contains a reference link to official documentation at https://cmux.com/pt-BR/docs/api. This is a legitimate vendor-owned resource for the cmux tool and does not involve automated downloads, package installations, or remote code execution.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 8, 2026, 12:26 AM
Security Audit — agent-trust-hub — cmux-cli