Powerful prompt engineering techniques for LLM hacking – how Large Language Models are hacked.
Posted inAI/Machine Learning

Powerful prompt engineering techniques for LLM hacking – how Large Language Models are hacked.

As Large Language Models (LLMs) such as ChatGPT, Perplexity, and Gemini become more prevalent, hackers are, of course, finding ingenious ways to hack them. They succeed because LLMs struggle to distinguish between legitimate instructions and adversarial inputs. Sure, they have some defenses, including input sanitization, output filtering, and adversarial training, but thus far, no foolproof solution exists to stop a hacker from tricking an LLM into doing what they want.
Washington Post breach leaks journalists’ email accounts and sensitive email messages.
Posted inCybersecurity News

Washington Post breach leaks journalists’ email accounts and sensitive email messages.

In mid-June 2025, The Washington Post disclosed a significant cyberattack targeting its email system, resulting in the compromise of several journalists’ Microsoft email accounts. The breach was discovered on a Thursday evening, and staff were notified via an internal memo on Sunday, June 15, 2025. The memo, signed by Executive Editor Matt Murray, described the breach as a “possible targeted unauthorized intrusion”.
NIST offers SP 1800-35 comprehensive guide to Zero Trust Architecture (ZTA) with 19 real-world examples.
Posted inCybersecurity News

NIST offers SP 1800-35 comprehensive guide to Zero Trust Architecture (ZTA) with 19 real-world examples.

NIST Special Publication 1800-35 (SP 1800-35) is a comprehensive guide developed by the National Institute of Standards and Technology (NIST) to help organizations implement a Zero Trust Architecture (ZTA) in modern enterprise environments. This publication is the result of collaborative work between NIST’s National Cybersecurity Center of Excellence (NCCoE) and 24 industry vendors, aimed at demonstrating end-to-end zero trust solutions using commercially available technologies.
Dark web’s Archetyp Market seized by international law enforcement
Posted inCybersecurity News

Dark web’s Archetyp Market seized by international law enforcement

Archetyp Market, one of the longest-running and most significant drug marketplaces on the dark web, was dismantled in a major international law enforcement operation this week. The platform had operated for over five years, facilitating the sale of illegal drugs—including fentanyl and other synthetic opioids—by thousands of vendors to hundreds of thousands of users worldwide.
More than 20 security issues, including five zero-day exploits, discovered in Salesforce Industry Cloud.
Posted inCybersecurity News

More than 20 security issues, including five zero-day exploits, discovered in Salesforce Industry Cloud.

AppOmni researchers identified over 20 security issues—including five zero-day vulnerabilities and numerous insecure configurations—in Salesforce Industry Cloud, particularly within its OmniStudio suite. These findings highlight the risks that can arise from default settings, low-code development shortcuts, and customer misconfigurations, potentially exposing sensitive business and personal data to unauthorized access.
Tenable releases patches for Tenable Nessus Agent fixing high-severity vulnerabilities that could allow users to escalate privileges and execute code.
Posted inCybersecurity News

Tenable releases patches for Tenable Nessus Agent fixing high-severity vulnerabilities that could allow users to escalate privileges and execute code.

Tenable recently released important security patches for the Nessus Agent, addressing several high-severity vulnerabilities that could allow local users to escalate privileges, overwrite or delete files, or execute arbitrary code with elevated privileges. The most recent fixes are included in Nessus Agent version 10.8.5, released in June 2025.
Zoomcar says cybersecurity breach has laked data on more than 8 million users.
Posted inCybersecurity News

Zoomcar says cybersecurity breach has laked data on more than 8 million users.

Zoomcar Holdings, a peer-to-peer car-sharing platform operating primarily in India and Southeast Asia, says it has realized a significant cybersecurity breach affecting approximately 8.4 million users. The breach was first identified on June 9, 2025, when several Zoomcar employees received external communications from a threat actor claiming unauthorized access to company systems.