The Skill installs dependencies and starts user-provided applications in an isolated ChatGPT execution environment. A project ZIP can contain malicious install scripts, build scripts, test code, or runtime behavior. Review the project source and use the Skill only in an environment intended for untrusted execution.
- production credentials or API keys;
- real customer or regulated data;
- payment credentials;
- live OAuth client secrets;
- production database dumps;
- private certificates or signing material.
Use test-only fixtures, disposable accounts, local emulators, and .env.example values.
The Skill may create or seed a local test account through repository-provided fixtures, a test API, or a disposable local database. It must not bypass live CAPTCHA, OTP, OAuth, payment, anti-abuse, or third-party security controls.
Reports may include source excerpts, local paths, request metadata, visible test data, screenshots, and patches. Review and sanitize all artifacts before sharing publicly.
Do not open a public issue containing exploit details, credentials, or private project code. Contact the repository owner through a private GitHub security advisory when available.