Welcome to g8e Discussions!

[Badoot]

👋 Welcome to g8e Discussions

We are building the governance layer between AI agents and production infrastructure. The core premise of this project is absolute: The AI reasons. You decide. The architecture enforces it.

If you are here, you likely share our fundamental paranoia about giving unconstrained LLMs write access to your fleet. This is the space to debate the protocol, share deployment strategies, and help us harden the foundation.

🛠 What belongs here

• Architecture & Threat Modeling: Tear apart our security model. We want rigorous debate on our mTLS implementation, the LFAA (Local-First Audit Architecture) vault encryption, and Sentinel’s MITRE ATT&CK-mapped threat detectors.
• Operator Extensions: Share how you are deploying g8e across your infrastructure. Discuss implementations for Cloud Operators (AWS Zero Standing Privileges), MCP gateway connections, or air-gapped environments.
• Refactoring the "Rough Edges": We are transparent that a significant portion of this codebase was built with AI assistance—which means there is technical debt, hallucinated logic, and areas needing cleanup. If you want to propose a massive refactor of the control plane or the Go operator, discuss it here before opening a PR.
• Engineering Q&A: Ask questions about FIDO2/WebAuthn setup, SQLite WAL constraints, or managing the internal platform CA.

🚦 Issues vs. Discussions

• Issues are for reproducible bugs, broken tests, and strictly defined feature implementations.
• Discussions are for architectural debates, threat modeling, and open-ended design questions.

---

To get started: Comment below with what infrastructure you are trying to govern, which LLM provider you are piping into the engine, and the specific failure modes you're most worried about.