🛡️ NOTABLE PROJECT: GUARDIAN-RUNTIME
A local runtime firewall for AI agents. Every LLM prompt and response is intercepted at the transport layer, inspected and governed entirely on-device, before any data leaves your machine.
┌─────────┐ ┌─────────────────────────────┐ ┌───────────────┐
│ Agent │ ─ prompt ─────>│ GUARDIAN-RUNTIME │ clean prompt ─>│ LLM Provider │
│ │ <─ safe reply ─│ * FinOps budget enforcer │ <─ response ───│ OpenAI/Gemini │
│ │ │ * Credential leak blocker │ │ /Claude │
│ │ │ * Token compression │ │ │
└─────────┘ └────── 100% on-device ───────┘ └───────────────┘
