nforum quietly stops logging every AI exchange to disk

A fork that was writing the raw output of every AI response to a file on the server - in production - now keeps that turned off by default.

securityinfrastructure

Until this change, nforum's fork saved the full raw stream of every AI response to a file on disk, no matter where the code was running - live production included. That's two problems at once. The file grows without limit, and, more to the point for legal work, nobody decided what ended up in it: whatever flowed through the model's responses got written down on the server.

nforum's fix switches that logging off unless an engineer deliberately turns it on to debug a problem, which is how this kind of logging is supposed to behave. It's a small change, but it closes a path where sensitive client material could quietly accumulate on disk that nobody was watching.

So what Anyone running a legal-AI tool should ask where the model's raw output gets written, and whether anyone meant for it to land there.

View this fork on GitHub →

Spotted something wrong? Or know the PR text has fresher detail than the writeup above?

Commits in this thread

2 commits from nforum/mike, oldest first. Source extracted verbatim from the harvested git log.

SHA Subject Author Date
95cf296f fix claude raw stream logging Eli Fayerman 2026-05-04 ↗ GitHub
08eae79a Merge PR #29: fix claude raw stream logging Bojan Plese 2026-05-07 ↗ GitHub

Capture this thread into my fork

Download a single Markdown prompt that tells Claude how to port every commit above into your working tree — adapting paths and structure to match your repo. Run it via claude -p < capture-thread-48.md from inside the repo you want the changes in.

⬇ Download capture-thread-48.md