We just open-sourced Paperclip: the orchestration layer for zero-human companies
It's everything you need to run an autonomous business: org charts, goal alignment, task ownership, budgets, agent templates
Just run `npx paperclipai onboard`
https://t.co/wuDdEmrSMx
More 👇
The Paperclip team merged/closed over 100 pull requests yesterday
We're not quite at Openclaw's pace, but we're accelerating
Keep the contributions coming!
Refactored the "skills store" inside Paperclip
Company-shared skills are one of the most important workflows and now it's easier to find / share / install new skills
Low-trust agents means you can have an agent review untrusted inputs and if he’s prompt-injected it won’t pwn the rest of your Paperclip org
(requires sandboxed instances)
every agent has to figure out where to "do the work"
the harnesses are trained on working with local text files but this becomes a problem when you run in a stateless sandbox
here's a thread of agent "disk-like" things i'm looking at:
@patrickc It’s bring-your-own agent - supports any harness
Has multi-user collaboration & VSC integration
The outputs are shared among users
You can setup both skills and routines for recurring work
We just open-sourced Paperclip: the orchestration layer for zero-human companies
It's everything you need to run an autonomous business: org charts, goal alignment, task ownership, budgets, agent templates
Just run `npx paperclipai onboard`
https://t.co/wuDdEmrSMx
More 👇
📎📎📎
It feels like it should be easy to loop over your task manager with an agent
but you immediately run into all sorts of problems that Paperclip solves
@dotta i tried doing my own linear setup months ago but realized I need better control of worktree isolation, steering, etc. - you’ve literally solved all of that beautifully. feels like the future