Search: “agent safety”

48 stories

Searchkeyword smart Sortlatest hot importance relevance

DomainAll Cyber AIWindow24h 3d 7d 30d AllIn the wild Important only Show duplicates

CategoryAll Vulnerability 144 AI safety & security 141 Policy & legal 79 Exploit / PoC 59 Ransomware 58 Malware 53 Data breach 47 Threat actor 39 Phishing & fraud 39 AI industry 36 Industry 35 Model release 18 AI policy 14 AI research 13

SeverityAll critical 263 high 107 medium 170 low 95 info 174

Claude Opus 4.6 improves agentic performance and model safety

Help Net SecurityFeb 6, 00:00 UTC · Feb 6, 2026Vulnerability

30

Microsoft open-sources tools for designing and testing AI agents

Help Net SecurityMay 21, 00:00 UTC · May 21, 2026Vulnerability1

30

Microsoft Expands Sentinel Into Agentic Security Platform With Unified Data Lake

The Hacker NewsOct 2, 04:51 UTC · Oct 2, 2025Vulnerability1

30

"Context bombs" can frustrate AI-driven attacks, researchers found

Help Net SecurityJul 14, 00:00 UTC · Jul 14, 2026Vulnerability

30

New BioShocking Attack Tricks AI Browsers Into Leaking User Credentials

The Hacker NewsJun 30, 08:43 UTC · Jun 30, 2026Vulnerability1

30

Meta’s Muse Spark takes AI a step closer to personal superintelligence

Help Net SecurityApr 17, 07:09 UTC · Apr 17, 2026Vulnerability

30

OWASP 2026 LLM Top 10: "The model will be fooled"

Help Net SecurityAug 6, 00:00 UTC · Aug 6, 2026Vulnerability

30

Agentic AI and the risks of unpredictable autonomy

Help Net SecuritySep 9, 09:51 UTC · Sep 9, 2025Vulnerability

30

Make OpenAI’s models misbehave and earn a reward

Help Net SecurityMar 27, 00:00 UTC · Mar 27, 2026Vulnerability

30

AWS, Google, and Vercel Agent Flaws Let Attackers Trigger Tools Without Running the Model

The Hacker NewsAug 6, 09:25 UTC · Aug 6, 2026Vulnerability CVE-2026-18830 CVE-2026-18236 CVE-2026-64650+1 CVEs1

35

The AI criminal mastermind is already hiring on gig platforms

Help Net SecurityApr 27, 00:00 UTC · Apr 27, 2026Vulnerability

30

Terra adds continuous network exploitation validation to its platform

Help Net SecurityMay 21, 00:00 UTC · May 21, 2026Vulnerability

30

GPT-5.6 gets better at cybersecurity

Help Net SecurityJun 29, 00:00 UTC · Jun 29, 2026Vulnerability

30

Google's Naptime Framework to Boost Vulnerability Research with AI

Infosecurity MagazineJun 25, 12:30 UTC · Jun 25, 2024Vulnerability

30

Hacker Runs Hermes AI Agent Unattended for Post

The Hacker NewsJul 24, 10:15 UTC · Jul 24, 2026Vulnerability CVE-2026-31431 CVE-2026-43284 CVE-2026-43500+2 CVEs

35

AI takes on a bigger role in finding Chrome vulnerabilities

Help Net SecurityJul 30, 00:00 UTC · Jul 30, 2026Vulnerability

30

Hadrian unveils Nova, enabling continuous, AI-powered offensive security testing

Help Net SecurityMar 24, 00:00 UTC · Mar 24, 2026Vulnerability

30

Single Prompt Enables ChatGPT to Execute Full Cyber

Infosecurity MagazineJul 16, 13:30 UTC · Jul 16, 2026Vulnerability

30

BlueCodeAgent helps developers secure AI-generated code

Help Net SecurityNov 20, 00:00 UTC · Nov 20, 2025Vulnerability1

30

Google’s Gemini 3.5 Flash Cyber becomes a vulnerability hunter

Help Net SecurityJul 22, 00:00 UTC · Jul 22, 2026Vulnerability

30

Deterministic + Agentic AI: The Architecture Exposure Validation Requires

The Hacker NewsApr 15, 00:00 UTC · Apr 15, 2026Vulnerability

30

CalypsoAI Security Leaderboard offers safety ranking of major GenAI models

Help Net SecurityFeb 27, 00:00 UTC · Feb 27, 2025Vulnerability

30

Week in review: 40 open-source tools securing the stack, invisible IT to be the next workplace priority

Help Net SecurityDec 14, 00:00 UTC · Dec 14, 2025Vulnerability1

30

Anthropic releases Claude Opus 4.7 with automated cybersecurity safeguards

Help Net SecurityApr 16, 00:00 UTC · Apr 16, 2026Vulnerability

30

Time-of-Check Time-of-Use Attacks Against LLMs

Schneier on SecuritySep 18, 11:06 UTC · Sep 18, 2025Vulnerability

30

Digital Car Keys Are Coming

Schneier on SecurityNov 28, 20:19 UTC · Nov 28, 2023Vulnerability

30

Frontier AI models collapse under multi-turn AI attacks, Cisco finds

Help Net SecurityJun 19, 11:19 UTC · Jun 19, 2026Vulnerability

30

Excessive agency in LLMs: The growing risk of unchecked autonomy

Help Net SecurityApr 8, 00:00 UTC · Apr 8, 2025Vulnerability

30

TuxBot v3 Evolution Shows Signs of LLM

The Hacker NewsJul 15, 00:00 UTC · Jul 15, 2026Vulnerability

30

NVIDIA’s Open Security AI Alliance Is Missing Some Big Names

Infosecurity MagazineJul 28, 09:40 UTC · Jul 28, 2026Vulnerability

30

We Scanned 1 Million Exposed AI Services. Here's How Bad the Security Actually Is

The Hacker NewsMay 5, 10:30 UTC · May 5, 2026Vulnerability

30

New Anthropic Fable 5 Is a

Infosecurity MagazineJun 10, 11:00 UTC · Jun 10, 2026Vulnerability

30

Qualys at RSA Conference: Implementing innovation

Help Net SecurityFeb 10, 00:00 UTC · Feb 10, 2017Vulnerability

30

BadPower: Fast chargers can be modified to damage mobile devices

Help Net SecurityJul 20, 00:00 UTC · Jul 20, 2020Vulnerability

30

Sysdig launches new version of its Cloud-Native Visibility and Security Platform

Help Net SecurityApr 10, 00:00 UTC · Apr 10, 2019Vulnerability

30

Friday Squid Blogging: Live Colossal Squid Filmed

Schneier on SecurityApr 18, 21:02 UTC · Apr 18, 2025Vulnerability1

30

DarkMoon: Open-source AI pentesting platform

Help Net SecurityJun 29, 00:00 UTC · Jun 29, 2026Vulnerability

30

ChatGPT's Data Protection Blind Spots and How Security Teams Can Solve Them

The Hacker NewsApr 20, 11:26 UTC · Apr 20, 2023Vulnerability

30

The Promptware Kill Chain

Schneier on SecurityMar 14, 11:18 UTC · Mar 14, 2026Vulnerability

30

Cybersecurity jobs available right now: July 7, 2026

Help Net SecurityAug 7, 10:57 UTC · Aug 7, 2026Vulnerability1

30