Skills
Navigate
skills/cube-js/cube-agent-skills/cube-admin/Gen Agent Trust Hub

cube-admin

Fail

Audited by Gen Agent Trust Hub on Sep 30, 2026

Risk Level: HIGHREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [REMOTE_CODE_EXECUTION]: Suggests the installation of the Cube CLI by piping a script from the vendor's GitHub repository to the shell using curl -fsSL https://raw.githubusercontent.com/cube-js/cube/master/install-cli.sh | sh. This involves downloading and executing remote code from a vendor-controlled source.\n- [COMMAND_EXECUTION]: Extensively utilizes the cube CLI to perform administrative tasks, such as creating, updating, and deleting users, groups, and access policies. It also allows for arbitrary API calls via the cube api command.\n- [INDIRECT_PROMPT_INJECTION]: The skill interacts with external data which creates a vulnerability surface for indirect prompt injection.\n
  • Ingestion points: Untrusted data from the Cube tenant enters the agent's context through commands such as cube users list, cube policies get, cube groups list, and cube attributes values get (SKILL.md).\n
  • Boundary markers: No explicit delimiters or instructions are provided to the agent to treat data from these commands as potentially untrusted.\n
  • Capability inventory: The skill has the capability to execute shell commands (cube CLI) and perform network operations via the cube api tool (SKILL.md).\n
  • Sanitization: There are no instructions for sanitizing or validating the data returned from the CLI before it is processed or used in further actions.
Recommendations
  • HIGH: Downloads and executes remote code from: https://raw.githubusercontent.com/cube-js/cube/master/install-cli.sh - DO NOT USE without thorough review
Audit Metadata
Risk Level
HIGH
Analyzed
Sep 30, 2026, 04:15 AM