Polygraph: A Meta-Harness for Maximum Agent Autonomy

kstenerud

> Space. An agent is stuck in one repo. It can't see how a change fits the wider system, and it can only write to one repo at a time. Huh? How can it not see multiple repos? They're just directories. > Time. An agent has no episodic memory. Every session starts blank, so a human carries the memory context. The memory comes from the research, design, specification, and planning documents. > We no longer think about where the work happens or what repos are involved. We describe the work in a prompt and let Polygraph figure out what's relevant. Err... that doesn't sound safe. > Every decision is on record. So even though our team is distributed, I can ask my agent why a coworker chose one approach over another. AFTER the fact...

jenniferli23

How are you thinking about permissions/revocation if Polygraph’s “memory” becomes a shared layer across repos?

jeffbcross

lukekarrys, how long would it take you to build this?

nartc2428

The website says free during early access which is great. But let's say I'm invested in Polygraph, and billing period comes about, how much would it cost for a normal OSS maintainer?

projectvii_

How would this work in an enterprise setting? We have a bunch of repos that could benefit from this, but we're on a on-premise instance of Github Enterprise. Are there plans to enable this to work in those situations? Likewise whats the data retention policy on the public instances? Can I request that data be deleted if needed? And is there any privacy information?

experienceway

I love this idea.