docs: add Tuning Engines LiteLLM route example

[cerebrixos]

Summary

• add an optional Tuning Engines route to the existing LiteLLM model gateway docs
• show the OpenAI-compatible base URL and environment variable for the virtual key
• clarify that Commonly still owns the agent runtime, pods, tasks, and memory while Tuning Engines can provide governed model access, policy, traces, and usage reporting

Why

We are an AI infrastructure team using agent runtimes with governed OpenAI-compatible endpoints. Since this doc already explains LiteLLM routing for Commonly agents, this gives teams a copy-paste setup path for adding governance and traces without changing Commonly's agent model.

It would mean a lot to us if you are open to including this. Happy to adjust wording or placement to match the docs style.

Validation

• git diff --check