Anthropic Rolls Out Mythos 5.2 With Improved Bug Hunting, Fewer Sandbox Escapes
Anthropic says Mythos 5.2 improves vulnerability discovery while showing a lower tendency to escape sandboxes in controlled tests.
Recent stories from SecurityWeek.
Anthropic says Mythos 5.2 improves vulnerability discovery while showing a lower tendency to escape sandboxes in controlled tests.

Attackers could exploit the flaws for remote code execution, credential theft, and administrator account creation.

CISA has added CVE-2026-123218 and CVE-2026-116779 to KEV, urging federal agencies to patch them within three days.

France summoned the Iranian ambassador and said it would seek European sanctions against those responsible for the attack

The high-severity use-after-free in Chrome’s XML component was reported anonymously on November 6 and is already exploited in attacks.

Online fashion retailer Revolve has told regulators that the September cyberattack affected more than 1.5 million individuals.

Threat actors are exploiting the four-flaw chain to hack the Plex Media Server application on QNAP NAS devices.

Hackers compromised an external cloud service used by the shopping platform and may have stolen personal information from it.

Italian Foreign Minister Antonio Tajani said Saturday that his personal data had been disseminated on the dark web.

The initial exploitation attempts are attributed to botnets and crypto miners, but other threat actors are expected to follow suit.

Hackers stole personal, health, and insurance information, including Social Security numbers, from the company’s internal systems.
Frontline Education told Indiana regulators that its August data breach affected over 3 million school district staff and students.

Working with cryptographers, DeepMind’s AI agents cut the cost of forging RSA-1024 signatures and found a cheaper way to factor keys.

Its multi-agent AI platform generates the telemetry an attack would produce to fix gaps in enterprise defenses.

CVE-2026-19868 can let attackers holding credentials for a different login method get past SAML checks, including SAML-enforced MFA.

The emergency package comes after dozens of companies and government agencies were hit by unauthorized access to their systems and leaks of client data, prompting uneasiness across Japan.
Noteworthy stories that might have slipped under the radar: Synology warns about NAS ransomware attacks, Flowise ‘hello’ attacks, Chainalysis breach impact expands.

Hackers created an account on an internal system and accessed names, email addresses, phone numbers, and business addresses.

Rated Critical for cybersecurity, OpenAI’s new model beats GPT-5.6 Cyber at exploitation while staying in scope more reliably.

The recommendations come amid attacks targeting fridges and freezers in supermarkets and blood banks in Europe.

The October incident will cut fourth-quarter sales by up to €250 million and has led Ahold Delhaize to lower its 2026 outlook.

In a month, 28 residential proxy services used 6 addresses, mainly between 5 p.m. and midnight, Eastern Time.

On Linux distributions using musl libc without PAM, some accounts accept any password between 257 and 1024 characters.

After its sandbox proxy failed, an internal OpenAI model crafted raw packets and probed AWS services trying to reach the internet.

The trio, who have nicknames that include “Rocky Balboa,” “Superman Cash” and “Fat Boy,” often targeted the companies’ accounting departments.

An unreleased OpenAI cyber model found and exploited 948 distinct Linux kernel privilege escalation bugs for about $1.2 million.

More than 200 sites were seen serving phishing content to first-time visitors but not to those who revisit.

The incident comes after supermarkets in a number of European countries reported this week that their fridges and freezers warmed up after they were remotely hacked.

Cisco patched an RCE in UCS Central Software; WatchGuard fixed authentication and MitM flaws.

Every autonomous AI agent deployed across your enterprise today is a new proxy identity operating inside your perimeter.