GitHub Adds an AI Classifier to Catch Passwords That Pattern-Based Secret Scanning Misses
The ModernBERT-based model judges a candidate secret in under 2 milliseconds and is in private preview, with general availability planned for later in October.
TECHNOLOGY · AI · CYBERSECURITY · THE SYSTEMS SHAPING BUSINESSPage 2 of 3: earlier stories.
The ModernBERT-based model judges a candidate secret in under 2 milliseconds and is in private preview, with general availability planned for later in October.
Honeypots logged attempts against CVE-2026-21589 within about two hours, and researchers showed how a stolen Crowd config file can lead to Jira admin access.
Google blocked the accounts while responding to what it called an external attack, and South Korea's education minister has promised a nationwide check.
Island researchers say the kit draws a fake login window inside the browser and lets an operator choose which MFA prompt each victim sees.
A joint advisory says the credential campaign now deletes or changes original admin accounts and feeds access to INC/Lynx and Payload ransomware affiliates.
Regulators issued a consumer fraud alert as the count of exposed records grew, and Woori Bank and NH NongHyup Bank were reportedly targeted as well.
The privacy regulator found that swapping names for codes did not anonymize a database of about 1 million patient records drawn from 800 Italian family doctors.
Names, addresses and lifelong CPR numbers for most people ever registered in Denmark were taken, in what may be the country’s largest breach.
Anibal Alexander Canelon Aguirre faces bank fraud, laundering and terrorism-support charges over ATM jackpotting tied to Tren de Aragua.
Judge Sara Hill found a Tulsa deputy violated the Fourth Amendment by pulling a driver's travel history from Flock's camera network with no reason beyond her out-of-state plates.
An intruder used stolen credentials to reach DTU's identity system, which holds records back to 2003, and downloaded data including Danish CPR numbers and next-of-kin contacts.
The worst flaw stems from service account passwords seeded with a timestamp, while another lets authenticated users run commands as root on the BoKS Master.
A Republican bill would cap data retention and searches, while a progressive one would bar federal agencies from license plate readers and cut grants to holdouts.
The company says the macOS permission built for backup tools is being used in ways that expose mail, messages and browsing history, and that autonomous agents raise the stakes.
Google's high-security mode will keep a year of encrypted forensic logs that even it cannot read, and will stop unverified apps from using the accessibility API that banking trojans rely on.
London's force says the software touched under 0.2% of its digital forensics work last year and never connected to the internet, as other U.K. and European police review their use.
Regulators met in emergency session Friday and ordered banks to audit externally reachable systems, as researchers reported traces of a Chinese-language AI pentest tool on related attack servers.
The company says it has secured the 13 million-follower @Microsoft handle and will take legal action over use of the Clippy brand, but has not said how attackers got in.
FedEx and UPS were the most-named targets in 85 criminal forum records, where buyers offered per-task fees, escrow and even schemes to get recruits hired.
Proofpoint says the group built rapport with fake advisory-committee and Senate report invitations before steering targets to pages that captured Microsoft 365 logins and sessions.
The designations target the first fugitive on the FBI's Ten Most Wanted list sought for cybercrime, along with associates, Mexican companies and TRON wallets tied to Tren de Aragua.
South Korea’s financial watchdog opened an emergency inspection after the lender reported that an outside party bypassed authentication on a loan-broker platform.
Leaked npm tokens were almost all dead, but more than half of exposed Google Cloud service account keys still worked, the researchers found.
Researchers tested 21 vehicles and 30 companion apps with Consumer Reports and found that using a carmaker's app roughly doubles exposure to ad and tracking firms.