Hermes Agent adds LM Studio support

This is a useful integration more than a flashy launch: it cuts friction for anyone already using LM Studio as a local inference host.

• –Hermes Agent can point at an OpenAI-compatible LM Studio server, so setup stays close to the standard local-LLM workflow.
• –The 64K context guidance matters because agent tool stacks eat tokens fast; without that, local runs get cramped quickly.
• –JIT loading and reasoning-effort support make the integration feel intentional, not just “it happens to work.”
• –For Windows users, the WSL2 and network-bridging details are the real operational hurdle that will decide whether this feels smooth or brittle.