really interesting work! i am curious how you handle rule configuration for different protocols such as Postgres or ssh.
Thanks for open-sourcing it under MIT.
The hard part with these routers is deciding the cheap model is "good enough" without already knowing the answer.RouteLLM trained classifiers for it, Not Diamond sells it as a service. Curious how you decide a subagent task is safe to send on-device?
Worth noting retention doesn't end at the model provider. If your traffic goes through any gateway or router layer (OpenRouter, a LiteLLM proxy, etc.) that layer sees every prompt too,