Latest cybersecurity reporting from selected sources.

Also, 23andMe Breach Settlement, GitHub AI Flaw, Ubiquiti Patches Critical BugsThis week, hidden breaches, researchers exposed a GitHub AI agent flaw, an Ubiquiti critical flaw, a ColdFusion flaw, a growing Chinese ORB, U.K. government FortiBleed exposure, 23andMe victims get $46.75 million, 3.8 million affected by Medtronic breach and Cerner's 2025 victim count went up.

AI Dependency Attack Reportedly Exposes Data and Source CodeA LiteLLM supply-chain compromise enabled attackers to harvest credentials and access internal environments at scale at Mercor. The firm was the first to confirm a LiteLLM breach, and researchers are warning about growing AI system exposure and limited visibility.

Cybersecurity Startup Exposed Lilli Using a Flaw as Old as the WebSecurity startup CodeWall disclosed this week that its autonomous AI agent breached McKinsey's internal AI platform Lilli in two hours on Feb. 28, accessing tens of millions of messages and hundreds of thousands of files through a basic, years-old database flaw.

Leaked non-human identities like API keys and tokens are becoming a major breach driver in cloud environments. Flare shows how exposed machine credentials quietly grant attackers long-term access to enterprise systems. [...]

The AI-assisted attack, which started with exposed credentials from public S3 buckets, rapidly achieved administrative privilges.

ChatGPT Maker Probes Third-Party Data Breach; OpenAI API Users' Information ExposedOpenAI has temporarily ceased use of Mixpanel after the analytics firm disclosed a breach affecting profile data of the artificial intelligence giant's API platform users. The company is notifying impacted organizations and watching for signs of data misuse.

OpenAI is notifying some ChatGPT API customers that limited identifying information was exposed following a breach at its third-party analytics provider Mixpanel. [...]

The breach may have exposed OpenAI API customers’ data

SonicWall has formally implicated state-sponsored threat actors as behind the September security breach that led to the unauthorized exposure of firewall configuration backup files

A single AI chatbot breach at Salesloft-Drift exposed data from 700+ companies, including security leaders. The attack shows how AI integrations expand risk, and why controls like IP allow-listing, token security, and monitoring are critical.

'Widespread Data Theft Campaign' Compromised Many Drift OAuth Tokens, Warn ExpertsThreat researchers report that "a widespread data theft campaign" traces to attackers stealing OAuth access tokens for applications integrated with Salesloft's AI chatbot Drift, then exfiltrating data. Victims include Salesforce customer Zscaler. Google Workspace instances were also breached.

Also: Ukrainian Hackers Find Evidence of Russian Child AbductionThis week, a Chinese duo arrested in Los Angeles for illegal artificial intelligence chip exports back to China, France extradited an accused Nigerian hacker, Ukraine hacked Crimean servers, Florida prison email leak, Tea App clone exposed users’ IDs.

AI acts like Pac-Man—devouring sensitive data across clouds, apps, and copilots. Varonis analyzed 1,000 orgs and found 99% have exposed data AI can access, exposing them to data risks. [...]

A threat actor claimed to get hold of an exposed GitLab configuration file containing Zendesk API access tokens

The data breach exposed more than 10m customer conversations from an AI call center platform in the Middle East

Also: US Fingers Russian GRU Hackers; Ohio City Sues ResearcherThis week, YubiKey 5 has a flaw, an Ohio city sued a researcher, the Irish regulator ended its GrokAI case, open-source AI tools exposed data, Starlink blocked X in Brazil, FCC banned Kaspersky, Intel addressed a researcher's claim, and Transport of London is still affected by a cyber incident.

Australian telco Optus allegedly left redundant website with poor access controls online for years The data breach at Australian telco Optus, which saw over nine million customers' personal information exposed, has been blamed on a coding error that broke API access controls, and was left in place for years.…

A threat actor, presumably from Tunisia, has been linked to a new campaign targeting exposed Jupyter Notebooks in a two-fold attempt to illicitly mine cryptocurrency and breach cloud environments

The breach was caused by a bug in an open-source library