TechPrivacy
Privacy · Page 2
Page 1- GMO Internet says it won't pay ransom after Onamae data leaks
The attacker posted data the company says is its own and its customers'. GMO urges people not to open it and says it has not confirmed ransomware.
- Honeypot captures 35 real shells using the PHP2Shell flaw
Remote code execution is now shown against PHP 8.4.26 with mod_php. Patches have been out since Oct. 21, and technical counts point to millions of hosts, not 100 million.
- Microsoft ties ConnectWise tenant attack to China-linked group
The 2,250 figure counts organizations the attackers tried to enter. How many were actually accessed is still unanswered.
- AI agents hit at least 15 Argentine health organizations
CrowdStrike says an operator automated the intrusions with Claude Code and took credentials and medical records from at least one victim.
- How a cyberattack heated up fridges at French supermarkets
Danfoss blames most cases on older controllers that no longer get security updates and were still reachable from the internet.
- Straiker: group tied to Mexico used Claude and Grok against banks
The security firm attributes to a group with alleged Mexican ties an AI system aimed at 400 financial institutions. No one else has confirmed it.
- 1Password lets an AI agent sign in without seeing the password
The feature is free on Individual and Family accounts. Muse and Grok Bot get it first; Claude Duet and ChatGPT Work have no date.
- QNAP has no patch yet for critical PHP flaws in its NAS devices
QTS, QuTS hero and other products are affected by PHP2Shell. QNAP sees exploitation attempts on the internet and gives no date for a fix.
- Judge orders Meta to pay $25 billion over data deception
New Mexico's attorney general calls it the largest state verdict against the company. Meta said it will appeal.
- Grok Bot terms leave the cost of its purchases to the user
Musk promised on X that SpaceX would cover purchase errors, but the agent's published terms leave that out and allow storing credentials.
- BeyondTrust flaw exploited since August now has a patch
CVE-2026-102764 hits Remote Support 26.2.3 and earlier; Google ties it to a China-linked group and CISA urges patching.
- CISA's deadline to fix the CVE-2026-21589 flaw expires today
Atlassian has patches for all eight Data Center products hit by an actively exploited critical flaw; installing them is up to each administrator.
- Mage OSS has a patch with no release for CVE-2026-111823
The critical flaw lets anyone create sessions without authenticating in versions through 0.9.80; the fix is in the code, but not in a published release.
- OpenAI and Anthropic keys may have leaked from OpenCode
The company cites unauthorized access to a legacy system and has not said how many accounts are affected or how the attacker got in.
- Meta will be able to crawl sites tied to business accounts
From Oct. 30, Meta's commercial terms let it crawl domains linked to an account and use them to train its products.
- ChatGPT lets advertisers target ads with emails and phone numbers
OpenAI documents that brands upload customer lists to include or exclude people. The ads appear only on the Free and Go plans.
- Google Messages will scan chats with strangers in the cloud
If users switch on "enhanced protection," Gemini will analyze conversations with senders outside their contacts on Google's servers.
- Fake Chrome app gets 100,000 downloads on Google Play
It was named Google Chrome and used its icon, but it came from another developer and worked as a chat. Google has removed it without explaining how it passed review.
- Cyberattack warms up fridges at French supermarkets
Attackers took over controllers exposed to the internet and triggered defrost cycles. Reports of affected stores range from 250 to 700 and no one has claimed the attack.
- Microsoft takes down YAM Store, a phishing kit market
With Health-ISAC, the company moved against three kits that, it says, enabled over a billion messages in 2026; it admits this is only part of the network.


















