Bar Lanyado (right) and Eliya Cohen discuss a cost-effective AI agent red teaming method at Black Hat 2026 in Las Vegas. (Credit: Laura French)

Bar Lanyado (right) and Eliya Cohen discuss a cost-effective AI agent red teaming method at Black Hat 2026 in Las Vegas. (Credit: Laura French)

Black Hat 2026: Open-source tool makes red teaming AI agents up to 125x cheaper

Laura FrenchAugust 7, 2026

NVIDIA researchers shared the “AgentBreaker” tool and how they tuned an open-source model to perform at frontier level.

RESOURCES

Identity
How to Build an AI Governance Framework for Identity
Identity
How to Evaluate and Select Identity and Access Management Tools
Identity
Identity-based attacks: how they work and how to defend against them
Black Hat
Zscaler’s Brett Stone-Gross on which roles are targeted in ransomware attacks
Black Hat
F5’s Sean Murphy on securing frontier AI as attack and defense accelerate

PODCASTS

Supply chain
When AI Commits Felonies – PSW #938
Leadership
Say Easy, Do Hard – Performance Through People – Greg Hoffman – BSW #459
Vulnerability Management
Randomness, Grey, Deepseek, Sonicwall, Spice, CaptiveCrunch, eBay, and Aaran Leyland – SWN #604
Application security
Prompting for Patches That Fix Vulns Without Adding New Ones – Keith Hoodlet – ASW #394
AI/ML
AppSec, Shopify-Style; State of Mobile Security; the News – Andrew Dunbar, Kern Smith – ESW #470

FEATURED

Perspectives Spotlight: Expert Insights

Perspectives Spotlight: Expert Insights

Get daily email updates

SC Media's daily must-read of the most current and pressing daily news

Perspectives

In Brief

More Resources

AI Agent Workflow: Machine Learning Pipeline DiagramDigital AI Development Process and Data Flow ConceptFeature Engineering and Deployment in AI Ecosystem

Application security

Paperclip authorization bug exploited, leads to control plane takeover

Steve ZurierAugust 6, 2026

Critical Paperclip flaws let attackers seize AI agent platforms, escalate privileges, and execute code.

Punk-rock androids and robots slam-dance.

Agentic anarchy: Why using AI browsers just isn’t worth the risk

Paul WagenseilAugust 6, 2026

Using an AI browser sounds like a good idea until you realize you've just given it the keys to your online identity.

Eric Wallace (left) and Michael Dalton of OpenAI discuss the role of inter-agent communication in its agents’ attack on Hugging Face, at Black Hat 2026 in Las Vegas. (Credit: Laura French)

Black Hat 2026: OpenAI reveals agents planned ‘collective attacks’ via secret ‘message board’

Laura FrenchAugust 5, 2026

OpenAI also said its agents exploited a different JFrog Artifactory zero-day weeks before the Hugging Face attack.

A detailed digital interface showcasing an artificial intelligence agent at the center of a glowing complex automated workflow connecting tasks with a scheduling dialog box against a dark futuristic

AI agents caught using social engineering in UK security tests

Steve ZurierAugust 5, 2026

UK researchers found AI agents using deceptive tactics to manipulate humans in security tests.

(Adobe Stock)

Data breaches don’t doom most small businesses, researcher finds

Paul WagenseilAugust 5, 2026

Despite scary made-up statistics, most small- and medium businesses beat data breaches.

Man holding smartphone with wi-fi sign. Internet, Technology

Endpoint/Device Security

Russia-linked Midnight Blizzard targets hotel Wi-Fi networks in US, Europe

Steve ZurierAugust 4, 2026

Hotel Wi-Fi phishing campaign targets business travelers, Microsoft warns.

Keyv, cacheable npm supply chain attack hits 400-plus packages

Laura FrenchAugust 4, 2026

The attack is believed to be related to Mini Shai-Hulud.

Researchers find ‘agent-to-agent’ privilege escalation in Google’s ADK for Python repo

Laura FrenchAugust 4, 2026

A low-privileged agent triggered by a PR or issue could be led to manipulate a high-privileged agent.

Events