Liability Doctrine | Sevorix | Terms of Service

Liability Doctrine

Last Updated: February 12, 2026

Welcome to Sevorix. By accessing our platform, using the Sevorix Lite CLI, or integrating our Context Firewall, you are entering into a binding agreement governed by this Liability Doctrine. If you do not agree to these terms, immediately cease all use of Sevorix services.

1. Nature of Service

Sevorix provides a Contextual Action Authorization Boundary (AAB). We provide the technical tools to intercept, inspect, and govern AI agent behavior. We do not provide legal advice, nor do we guarantee that our tools will stop 100% of malicious intents or agentic hallucinations.

2. Shared Responsibility Model

Sevorix is a security tool, not a replacement for human oversight. You are solely responsible for the configuration of your "Approved Intent" libraries and the final execution permissions granted to your agents. Sevorix is not liable for damages resulting from incorrectly configured policy sets.

3. License & Use

We grant you a limited, non-transferable license to use Sevorix tools for internal security purposes. You agree not to:

  • Reverse engineer the Faramesh-based detection engines.
  • Use Sevorix to facilitate the creation of malicious autonomous agents.
  • Attempt to bypass the Action Authorization Boundary in production environments without explicit testing authorization.

4. Limitation of Liability

TO THE MAXIMUM EXTENT PERMITTED BY LAW, SEVORIX INC. SHALL NOT BE LIABLE FOR ANY INDIRECT, INCIDENTAL, OR CONSEQUENTIAL DAMAGES, INCLUDING DATA LOSS, PRODUCTION OUTAGES, OR FINANCIAL LOSS RESULTING FROM AGENTIC BEHAVIOR, WHETHER OR NOT SUCH AGENTS WERE WRAPPED BY SEVORIX LITE.

5. Intellectual Property

All proprietary algorithms, "Intent Hashes," and the Faramesh-inspired execution control plane remain the exclusive property of Sevorix Inc. Your data and policy sets remain your property, subject to the anonymized metadata contribution outlined in our Privacy Doctrine.