Repository navigation
Are there plans to support augmentation code? #134
Description
Activity
Hi @poplovexzjip — thanks for raising this. I did a deep dive on Augment Code and compared it with the devin.cursorrules + Cursor stack. The key take-away is: Augment Code is positioned as a enterprise-grade “AI teammate”. In practice it behaves more like an opinionated platform than a pluggable tool.
That heft brings a lot of overlap with what we’re already building here—especially around long-term “Memories”, multi-step agent plans, and diff-style batch edits. Given those points, supporting Augment Code directly doesn’t buy us enough new surface area to justify the maintenance cost, and it would dilute the “open & hackable” ethos that makes this repo useful.
Happy to revisit if the ecosystem shifts or if someone wants to champion an external adapter, but that’s the status quo for 0.7.x.
Thanks again for the nudge—and for making sure we’re not rebuilding wheels we don’t have to.
- I did see augmentation code on SWE-Bench, which is kinda weird
- Soon there will be a universal file for all IDEs to use Directory support agentsmd/agents.md#9
I found that augmentation code is stronger than CURSOR WINDSUR