Albin Pollack

I ve spent the last few months obsessed with a specific failure in multi-agent systems: The Black Box Dilemma.

Most MAS setups I see (including my own v1) treat orchestration as "glue code." You have a central manager, a few agents, and a lot of hidden logic. When it fails, you re left guessing. As someone on IndieHackers put it: "You re just parallelizing chaos."

The Shift: From "Dictator" to "Blackboard"

In our latest iteration, we deleted the 430-line monolithic controller. Instead, we implemented a Blackboard Pattern.

Hey Product Hunt community!

I'm a solo maker from Sweden, currently building agentfarm.se. I ve spent the last few months exploring how to make AI agents more accessible and secure for businesses.

I'm a big fan of rapid iteration (yay vibe coding!), but I'm now focusing on the infrastructure needed to actually run these agents reliably and locally. To be honest, I don't have all the answers yet. I see this as a collective exploration into how we ll actually work with AI in the near future.

I m launching tomorrow (Tuesday!), but I m looking for more than just upvotes. I m looking for beta operators curious builders who want to test the limits of these agents with me and help shape the roadmap.

Just a few days to go

ZeroThreat is getting ready to introduce ZeroThreat 3.0: Agentic AI Pentesting, a new layer designed to move beyond vulnerability detection into real exploit validation. 

This isn t just another automated scan.