From the archive
Saturday, 21 November
Folded from 19 sources that day.
Sorted by quality and by how many sources covered it, the same for everyone.
-
GnuPG Warns Critical GPGME Flaw Could Expose App Data via Malicious OpenPGP Keys
CVE-2026-129975 is a critical memory bug in GPGME triggered by an OpenPGP key with an unusually long user ID. Versions before 2.3.0 are affected, and fixed releases are out.
-
AI Agents Found a Zero-Day in MONAI Label and Gave Attackers Root on a University Server
Amazon Threat Intelligence says suspected China-linked hackers used autonomous AI agents to find an unknown flaw in the MONAI Label imaging server. It lets unauthenticated users write files outside the storage folder.
-
Relevance Is Not Permission: Filter-First Vector Search in PostgreSQL
Shows how to enforce access control for RAG retrieval inside PostgreSQL using Row Level Security on a pgvector chunk table, so agents cannot bypass permission checks in application code.
-
CTO of 'RansomwareHelp' Indicted for Wire Fraud
An FBI sting showed that RansomwareHelp paid ransomware groups for decryption keys and marked up the price to victims, and its co-founder has been charged with wire fraud.
-
GitHub’s new “Security Advisory Permissions” feature is a massive win for OSS security sustainability
GitHub's Security Advisory Permissions lets teams and collaborators view, edit and publish advisories, not just admins. The Python Security Response Team tested it in private beta and found triage much easier.