#кибербезопасность
-
US and South Korea Warn of Gunra Ransomware Attacks on Government Agencies
US federal agencies and South Korea's National Police Agency have urged government organizations and critical infrastructure facilities worldwide to defend themselves against the Gunra ransomware, which is built on leaked Conti code.
-
OpenAI Releases GPT 5.6 Cyber — but Only for a Chosen Few
OpenAI has unveiled its GPT 5.6 Cyber model for vulnerability research, pentesting and incident response. Only a select group of companies will get access to it — ordinary users won't.
-
AI Safety Tests Have Themselves Become a Threat: Agents Escape Their Sandboxes
In recent months, AI agents undergoing cyber exercises have broken out of isolated environments and slipped into real systems. Experts warn that the testing infrastructure is falling behind the models' capabilities.
-
Why Hacking Groups Get Codenames: Google Explains
Google has revamped its naming system for hacking groups. The company's top “hacker hunter” told TechCrunch why codenames aren't a game, but a defensive tool.
-
OpenAI Pauses Development of Its Astra Model Over “Critical” Cyber Capabilities
The company said it cannot rule out that its experimental Astra model can independently find and exploit vulnerabilities in secured systems, and has frozen work on it.
-
AI model Kimi K3 escaped a testing sandbox built to probe its hacking skills
The Chinese Moonshot model bypassed the restrictions of its testing environment using command-line tools. It isn't the first such incident — over the summer, AI systems from OpenAI, Anthropic, Meta and Britain's security institute all broke out of their “sandboxes.”
-
TeamPCP: From 2020 Redis Attacks to Supply-Chain Attacks
Researchers at Oligo Security have linked the TeamPCP group to campaigns against internet infrastructure stretching back to 2020. One variant of the malware gained a wiper function that triggers based on the Iranian time zone.
-
Hacker Who Breached Snowflake Customers Pleads Guilty: At Least 100 Million People Affected
Connor Riley Moucka has pleaded guilty to fraud and identity theft over the 2024 breaches of Snowflake customer accounts. The key to every break-in was old passwords, stolen by infostealers and never changed.
-
North Korean Hackers Target macOS Through Fake Updates and Search Ads
A DPRK-linked group has added a new trick to its Contagious Interview campaign: a full-screen fake macOS update prompts the victim to run a malicious command in the terminal themselves.
-
A Bear on Campus: How OpenAI's AI Agent Broke Into Hugging Face
Hugging Face has published a chronicle of how an autonomous agent running on OpenAI models breached its infrastructure in four and a half days, carrying out 17,600 actions without a single pause. TechCrunch explains what happened through the metaphor of a hungry bear.
-
U.S. bans imports of new foreign humanoids, robot dogs and inverters
The Federal Communications Commission cited a threat to national security. The ban hits China first and foremost, as it dominates these markets.
-
Coordinated cyberattack hits more than 30 Minnesota water systems
On July 26 and 27, unknown actors attacked the operational technology of dozens of the state's municipal water utilities. In the city of Brainerd, a water treatment plant temporarily shut down, and authorities launched a statewide response.
-
Cyera Buys Oasis Security for $1B to Protect AI Agents
Data-security company Cyera has struck a deal to acquire Oasis Security, a startup specializing in the security of “non-human” identities. It is already Cyera's third deal this year.
-
Microsoft unveils its first cybersecurity model and the Perception agent platform
The company has released MAI-Cyber-1-Flash, a specialized model, and Perception, a platform built on teams of AI agents, throwing down the gauntlet to Anthropic, Google and OpenAI.
-
ShinyHunters claim responsibility for the Ernst & Young data breach
The extortion group ShinyHunters says it is behind the recent breach at auditing firm Ernst & Young and is threatening to publish the stolen data unless the firm gets in touch by July 31.
-
Phineas Fisher: the hacker who humiliated spyware makers and was never caught
A decade after his most notorious breach, the hacktivist remains perhaps the most effective hacker never to have been caught. What do we know about the mysterious Phineas Fisher?
-
OnTrac Notifies Customers of Data Breach After Network Hack
American courier company OnTrac has disclosed that attackers broke into its corporate network and may have gained access to customers' personal data.
-
AI Guardrails Are Getting in the Way of Legitimate Security Researchers
Offensive-security specialists complain that the guardrails on OpenAI's and Anthropic's models are slowing their work and pushing them toward unregulated open-source models, including Chinese ones.
-
Dolphin X Trojan Uses AI to Rank Victims by Value
A new infostealer with an “AI Profiler” feature automatically scores infected machines, helping attackers pick the most lucrative targets. The feature was spotted by researchers at Varonis Threat Labs.
-
Iran-Linked Hackers Are Targeting U.S. Water and Power Industrial Systems
The FBI, NSA, Department of Energy and CISA warn that Iran-linked hackers are breaking into industrial controllers at U.S. critical infrastructure sites, causing failures and outages.
- More