Harness Remote and Token Harness

Hi,
I've created and currently maintain two GitHub repositories focused on AI-related challenges I'm particularly interested in:

• Harness Remote – A native-session control plane for Codex, Claude Code, OpenCode, DeepSeek, Pi, and other coding agents and harnesses. It allows you to run, resume, and seamlessly hand off coding sessions across different machines, agents, and harnesses.
• Token Harness – A toolkit that collects, installs, and maintains some of the most effective technologies for reducing token consumption and coding agent costs, while measuring and tracking actual savings.

Both projects came from challenges I've personally encountered while working extensively with AI coding agents.

There's also something I'm particularly curious about that I want to ask you: how are you managing the rise of "vibe coders" within your organization?

I'm referring especially to people with deep domain expertise who aren't necessarily professional developers, but who can now build their own tools and applications thanks to AI.

How do you identify and empower these people, while making sure their solutions don't remain isolated experiments? Do you have a framework to turn their individual initiatives into reusable, maintainable, company-wide assets? And how do you balance their autonomy with IT governance, security, and architectural consistency?

I'd be really interested to hear how you're approaching this, both from a technological and an organizational perspective.