> Postgres makes those files searchable. memory.entries stores chunks and line references, memory.embeddings holds 384-dimensional vectors, and memory.claims tracks evidence, confidence, and status. 

Is each Muse instance running it's own Postgres??

That seems wildly wasteful, especially since earlier in the article it states that the Muse instance has a SQLite database and schema already...

I have no idea why people would ever want to touch anything from Meta.

Am I missing something? This isn't a vulnerability. Your agent can see the files in its virtual environment. SSH keys are also not necessarily confidential. Please don't use AI to write blog posts.

I assume SOUL.md was empty.

Seriously, I want to know what's in there!

The internals are not _that_ reveling, most agents run a similar setup. Metas' responds is the most interesting here.

This isn't a bug and doesn't deserves any bounty. Each user gets isolated VM and that is the design and agent is able to access everything.

original post on x: https://x.com/heypeterjames/status/2102183576418558269

muse is a pretty capable agent but still asks for too many approvals to do tasks. I'm a student and have been going between muse and instinct

Everything about this is just embarassing

The tldr is that muse is heavily inspired by openclaw and should be considered FB’s version of it.