Coding agent progress:
We're through the core feature set, and most of the hardening.
- - All of the Ollama cloud models (except for Gemma4) are certified.
- - All of the Gpt Codex models are certified.
- - Hooks have been implemented.
- - Scheduler system has had a major upgrade.
- - Quality of life addons and tweaks have been added. We now have an orange progress chart that tracks shadow repos per project. Fun fun.
- - Slash commands fixed, upgraded.
- - Doctor and installer are much more user friendly now.
- - Skills are now standard. Par with Claude/Codex/Openclaw
- - LSP system is fully fleshed out, fully tested, and operational
- - Broken multi-range features removed (we don't need them for current models anyway)
- - Mission based sub agent system is fully debugged, and all shiny.
Currently
~ Working on decoupling roles/jobs/personas. That's going to make the system a lot more customizable.
~ Observability layer. So you can watch 20 instances work in one UI.
~ Discord and Slack are in flight. Still a couple of problems to be solved there that I need to work through.
This is really starting to look like a finished product.