ask-ui

Warn

Audited by Gen Agent Trust Hub on Aug 17, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDATA_EXFILTRATION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses child_process.spawn in scripts/adapters/process-utils.mjs to execute external CLI tools (claude, codex, powershell.exe). While these are intended for integration with specific AI agent environments, the spawnCli function enables the execution of these commands with user-influenced arguments and context.
  • [EXTERNAL_DOWNLOADS]: The assets/app/index.html file includes a script tag pointing to https://cdn.bootcdn.net/ajax/libs/tailwindcss-browser/4.1.13/index.global.min.js. This is a third-party CDN resource loaded into the browser environment where session tokens and answers are handled.
  • [DATA_EXFILTRATION]: The skill implements a 'wake' mechanism in scripts/ask-ui.mjs and its adapters that sends session data, question paths, and answer paths to external CLI tools to resume workflows. While this is the core feature of the skill, it involves moving local session data across process boundaries.
  • [COMMAND_EXECUTION]: The skill uses rundll32.exe, open, or xdg-open to automatically launch the system browser to a local URL containing a sensitive access token.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Aug 17, 2026, 02:21 AM
Security Audit — agent-trust-hub — ask-ui