Biztonsági szemle

The company, one of four finalists in this year's Black Hat USA Startup Spotlight competition, uses multi-agent system to build AI Digital Employees.

We uncovered 1,100+ exposed Ollama LLM servers—20% with open models—revealing critical security gaps and the need for better LLM threat monitoring.

A user reported a bug in pdf-parser: when dumping all filtered streams, an error would occur:

Wireshark release 4.4.9 fixes 5 bugs.

Auditing must be seen for what it truly can be: a multiplier of trust, not a bottleneck of progress.

Generating exploits with AI and large language models shrinks the time to target software flaws, giving security teams scant time to patch. Can enterprises adapt?

A 2025. 35. hetére vonatkozó hírválogatás, amely az NBSZ NKI által 2025.08.22. és 2025.08.28. között kezelt incidensek statisztikai adatait is tartalmazza.

(c) SANS Internet Storm Center. https://isc.sans.edu Creative Commons Attribution-Noncommercial 3.0 United States License.

Three federal agencies were parties to a global security advisory this week warning about the extensive threat posed by Chinese nation-state actors targeting network devices.

The credit reporting agency said the breach was "limited to specific data elements" and didn't include credit reports or core credit information.

Flashpoint published its 2025 midyear ransomware report that highlighted the top five most prolific groups currently in operation.