Same reason everyone rolled their own auth in 2010, the problem is simple enough to DIY badly, complex enough that no standard fits everyone. My Claude Code needs SSH access but not rm. Your agent needs filesystem writes but not network. There's no "OAuth for syscalls" yet.
Comments
Same reason everyone rolled their own auth in 2010, the problem is simple enough to DIY badly, complex enough that no standard fits everyone. My Claude Code needs SSH access but not rm. Your agent needs filesystem writes but not network. There's no "OAuth for syscalls" yet.
This exists today in OSes in form of discretionary/mandatory permissions (e.g., SELinux, AppArmor, LandLocked).
Yea, but that's not click button imported from clerk oauth easy
this is the most insightful comment I've heard on this in a while
To me, OCI seems the best foundation to build on. It has the features, is widely disseminated, and we have a lot of practice and tooling already