I got jealous of the frontier labs and their agent swarms... so I've been playing with creating my own on a much smaller scale.
Today I'm open sourcing agent-horde, it's a coordination protocol for small agent swarms, and the pi extension that runs it. I use this heavily in my personal projects including for security research and OSINT investigations.
It works remarkably well and allows for someone with a low budget to observe how swarms coordinate at a small scale and where they fail.
Users can launch a horde of up to 12 agents that coordinate on a task (horde of hordes coming soon). The agents coordinate on a markdown based ledger that is browsable via a human readable dashboard. Every run records an interactive timeline of events.
https://t.co/4VXIohnMl7
v0.20 implements real swarms AKA "horde of hordes" - you can now run agent-horde with something like 3/4 [prompt] and it will spin up 3 hordes of 4 members each, and partition the task appropriately.
Release notes here:
https://t.co/nysp9w7sYe
Point it at your favorite repo and ask it to do a security review.
I got jealous of the frontier labs and their agent swarms... so I've been playing with creating my own on a much smaller scale.
Today I'm open sourcing agent-horde, it's a coordination protocol for small agent swarms, and the pi extension that runs it. I use this heavily in my personal projects including for security research and OSINT investigations.
It works remarkably well and allows for someone with a low budget to observe how swarms coordinate at a small scale and where they fail.
Users can launch a horde of up to 12 agents that coordinate on a task (horde of hordes coming soon). The agents coordinate on a markdown based ledger that is browsable via a human readable dashboard. Every run records an interactive timeline of events.
https://t.co/4VXIohnMl7