Hermes Agent is in use now, and so far it feels pretty smooth: - One-click import of OpenClaw configurations and memories from the same server (though I didn’t do this myself) - Had it fix a Telegram communication bug, and after resolving it, it automatically generated a related Skill - The GPT5.4 model actually feels better in Hermes compared to OpenClaw… - Some details are nicely done, like Token burn progress and execution specifics. Hurry up and surpass the Claude Code experience… I’m going to tinker with it more and hope to uncover more surprises.