Terms of Use

Sentry observes selected file, subprocess, and network actions, evaluates those actions against a policy, and presents session signals in its dashboard. It is a developer tool, not a managed security service.

The project is provided for development, demonstration, and evaluation. You are responsible for deciding whether its controls are appropriate for your environment.

Only run Sentry against systems, codebases, credentials, and services that you are authorized to monitor.

• - Review generated policies before using them in an important environment.
• - Keep API keys, secrets, and production credentials out of agent prompts and demo data.
• - Use independent access controls, backups, and code review; Sentry is not a substitute for them.

Optional prediction and policy-evaluation features can call OpenAI APIs when you supply an API key. Your use of those APIs is governed by the applicable OpenAI terms and settings.

The project is supplied as-is, without warranties of any kind. Detection signals and policy decisions can be incomplete, incorrect, or unavailable, and must not be relied on as the sole protection for security-critical work.