Skip to content

fix(chat): verify TLS peers and fetch gpt4all.io over HTTPS - #3696

Open
SashaMIT wants to merge 1 commit into
nomic-ai:mainfrom
SashaMIT:fix/tls-verify-model-downloads
Open

fix(chat): verify TLS peers and fetch gpt4all.io over HTTPS#3696
SashaMIT wants to merge 1 commit into
nomic-ai:mainfrom
SashaMIT:fix/tls-verify-model-downloads

Conversation

@SashaMIT

@SashaMIT SashaMIT commented Aug 8, 2026

Copy link
Copy Markdown

Summary

  • Model catalog / release / news / default GGUF downloads used cleartext http://gpt4all.io/....
  • Those requests (and chat ingest / Mixpanel / health) forced QSslSocket::VerifyNone, disabling certificate verification.
  • Because catalog entries supply both download URL and hash, a MITM that poisons the catalog can make the post-download hash check accept attacker-controlled weights.
  • Switch gpt4all.io URLs to HTTPS and remove VerifyNone overrides so Qt uses default peer verification.

Test plan

  • Build gpt4all-chat and confirm model list / download still works against gpt4all.io
  • Confirm no remaining VerifyNone or http://gpt4all.io in gpt4all-chat/src

Made with Cursor

Remove QSslSocket::VerifyNone on model catalog/downloads/telemetry and
switch cleartext http://gpt4all.io URLs to https so MITM cannot poison
catalog hashes or executable GGUF weights.

Signed-off-by: Sasha Mitchell <sash.t.mitchell@gmail.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant