quick-voice

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 0.90). The code intentionally exposes server-side tools (read_file, file serving, run_bash) and forwards an OpenAI API key/ephemeral tokens, enabling a remote realtime model to read arbitrary local files, execute shell commands, and thereby exfiltrate secrets or perform remote code execution if misused.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.80). The launcher (scripts/launch.js) runs npm install at runtime, which fetches and installs packages from the npm registry (https://registry.npmjs.org/) and therefore pulls and enables execution of remote code required for the skill.