Security Weekly (Download Only)

Most enterprise AI today is a conversation — it summarizes, suggests, recommends. Harman unpacks what changes when AI actually executes across endpoints, and the governance problem that creates. Where does the human stay in the loop, and where do they get out of the way?

This segment is sponsored by Tanium. Visit https://securityweekly.com/tanium to learn more about them!

Show Notes: https://securityweekly.com/swn-611

Rob Allen from ThreatLocker joins us to discuss securing agentic AI with zero-trust controls, least privilege, and access controls to limit what agents can access and do.

This segment is sponsored by ThreatLocker. Visit https://securityweekly.com/threatlocker to learn more about them!

In the security news this week:

• Sixteen-year-old Linux LPEs still work
• Ubiquiti UniFi, patch it, also light on details
• If you remember magicJack, you too are old
• Slovakia doesn't trust its own speed cameras
• More homework on NIST's vulnerability database
• Your webcam, mic, and key light, all owned
• Printer moonlights as Minecraft server
• Zombie credit cards
• Your car's infotainment system fuels botnets
• Feds warn about AI-powered PLC attacks
• Can an AI actually reverse engineer its way out?
• Denver International's security breach, volume six
• Charlotte's breach and a parking company
• Why your ancient tech might be the safe one
• Microsoft counts billions of phishing emails
• A password vault that leaked to any website
• Australia sells password books at the post office
• Another perfect ten, this time in Entra ID
• Cisco's bug scores read like Olympic gymnastics

Show Notes: https://securityweekly.com/psw-941

The threat landscape has become more interconnected, disruptive, and complex. Ransomware is now as much about extortion and data theft as it is about encryption. Supply chain events can create outages that ripple far beyond the initial target, and business interruption increasingly comes from third-party and cloud dependencies. How should CISOs prepare for these scenarios?

Dan Bowden, Global Business CISO at Marsh, joins Business Security Weekly to discuss how to connect cyber risk to the outcomes boards care about most: resilience, financial exposure, regulatory impact, and reputation. CISOs need to position cyber as an enterprise risk, not a technical risk, that can be measured, prioritized, and managed alongside other strategic risks. Dan will discuss the results of Marsh's Cyber Catalyst research and Global Cyber Claims Report.

Visit https://securityweekly.com/infosecworld2026 and save 30% on your ISW pass with code: ISW26-SWSAVINGS

The Agent Is the New Insider: Why Human Risk Doesn't Stop at People: Black Hat Interview with Leslie Nielson, CISO at Mimecast

AI agents now act with the same credentials and access as the humans who deployed them, but without the judgment or accountability that comes with actual employment. Mimecast CISO Leslie Nielsen argues that treating agentic AI as a brand new, standalone security category is the wrong instinct: agents are an extension of human risk, and the controls organizations already use to manage people are the right foundation for managing machines. In this conversation, Nielsen unpacks the growing gap between security leaders who expect AI driven attacks and those who feel prepared for them, and what that gap means for CISOs walking the floor at Black Hat.

Segment Resources:

Mimecast's new whitepaper Securing The Agentic Enterprise: https://assets.mimecast.com/api/public/content/securing-the-agentic-enterprise?v=ea66db05 Mimecast's landing page for thought leadership resources: https://www.workprotected.com/ Mimecast's State of Human Risk Report: https://www.mimecast.com/resources/ebooks/state-of-human-risk/ Mimecast's Threat Intelligence Hub: https://www.mimecast.com/threat-intelligence-hub/

For more information about Mimecast please visit: https://securityweekly.com/mimecastbh

Ransomware Moves up the Org Chart: Managers Are Prime Targets: Black Hat Interview with Brett Stone-Gross, Sr. Director, Threat Intelligence at Zscaler

When a ransomware attack makes headlines, attention usually turns to the organization that was breached, the systems encrypted, data stolen, and disruption or ransom demand that followed. Less, if anything, is revealed about the employees compromised at the start of the attack, and what makes those individuals valuable targets.

New Zscaler ThreatLabz research examines this early stage of a real-world ransomware attack. ThreatLabz identified victims of a campaign associated with a ransomware group known for gaining initial access, stealing large amounts of corporate data, and selectively encrypting critical systems. The findings show who those victims were and how their roles and authority could help an attacker move deeper into an organization.

This is part of ongoing ransomware research by ThreatLabz. The Zscaler ThreatLabz 2026 Ransomware Report, coming in the next two months, will include additional data on ransomware victims, the latest ransomware trends, targets, and tactics, and the risks enterprises should prepare for next.

This segment is sponsored by Zscaler. Visit https://securityweekly.com/zscalerbh to learn more about them!

Show Notes: https://securityweekly.com/bsw-462

Fibonacci and the Unhappy Number, Hidden Sounds, Teams, Zimbra, Entra-ID, z.ai, Schrödinger's battery, Aaran Leyland, and More on the Security Weekly News.

Show Notes: https://securityweekly.com/swn-610

Sandboxing, least privilege, and monitoring are well-established controls in terms of the defenses they provide against unexpected and unauthorized actions. But being well-established in theory doesn't always translate to successful in practice. Kieran Human talks about some of the properties that a good sandbox should have and how monitoring creates a feedback loop to refine allow lists and access controls. In practice, the potential unpredictable behavior of an agent isn't much different from malware. We talk through some of the ways orgs can securely deploy agents without unnecessarily increasing their attack surface.

Resources

• https://www.threatlocker.com/blog/the-principle-of-least-privilege-for-ai-agents
• https://www.threatlocker.com/blog/applying-threatlocker-to-agentic-ai-tools
• https://www.threatlocker.com/blog/why-the-five-eyes-alliance-sees-zero-trust-as-the-best-defense-against-agentic-ai-threats

This interview is sponsored by ThreatLocker. Visit https://securityweekly.com/threatlocker to learn more about them!

Show Notes: https://securityweekly.com/asw-397

Interview with Rob Allen from Threatlocker

Safely enabling agentic AI for Businesses

OpenClaw was the wakeup call and businesses wanted to know how to block it. "Easy," Rob Allen said, "it's already blocked if you're using Threatlocker." Now that things have settled down a bit, those same businesses want to allow their employees to experiment with agents. We discuss how they can do it safely.

This segment is sponsored by ThreatLocker. Visit https://securityweekly.com/threatlocker to learn more about them!

Topic Segment

For this week's topic segment, we're discussing AI pentesting agents and how likely they are to get you into big legal trouble. You came home from Black Hat with a new, shiny AI pentesting agent. How can you be sure it isn't hacking the wrong company?

News Segment

Finally, in the enterprise security news,

1. we check the vibes
2. New MCP standard and AI text watermarking
3. what does combatting "cyber-enabled crime" mean?
4. A closer look at Cl0p
5. One 3rd party was responsible for all the AI sandbox escapes and hacking
6. reports
7. vulnerabilities
8. Comcast can track your movements with WiFi
9. A novel solution to the AI datacenter water use concerns

All that and more, on this episode of Enterprise Security Weekly.

Show Notes: https://securityweekly.com/esw-473

Surveillance, Murder Hornets, Portmantau, TrueCONF, Siemens, N-Able, Robo-Tips, and More on this episode of the Security Weekly News.

Show Notes: https://securityweekly.com/swn-609

In the security news this week:

• Cursor opens your repo, the repo opens you
• If you want the good model I'm going to need to see your ID
• Flock's a Flocking mess
• Defender was supposed to be the chosen one
• Side stepping Secure boot - twice
• SonicWall: a LAMP stack in a fancy case
• Macs don't get viruses, part infinity
• Flipper One, but why not Nix?
• NetScaler is back in the room
• Borrowing phone's good reputation
• USB and how to make Windows download stuff
• A KVM with the expensive letters removed
• Five steps to stop the webcam creeps
• PlexTrac acquired
• NIST asks the internet to fix the NVD
• Poland's health software has a very bad week
• If Apple pings you about spyware, believe it
• A macOS stealer that drives your browser for you
• T-Mobile's incident response tool of choice may suprise you, or not...

Show Notes: https://securityweekly.com/psw-940

Artificial intelligence has quickly evolved from a productivity tool into an active participant in many organizations' daily operations. As organizations give AI greater autonomy within their environment, they're also granting them access to sensitive systems and data. That creates a new challenge for IT and security teams: How do you enable AI to assist productivity without compromising security?

Rob Allen, Chief Product Officer at ThreatLocker, joins Business Security Weekly to discuss how zero trust principles can prevent an AI breakout. Rather than relying solely on the AI tool's built-in safeguards, organizations can choose to enforce security policies using ThreatLocker. Rob will discuss how ThreatLocker can enforce AI boundaries through Allowlisting, Ringfencing™, Endpoint Firewall, and Web Content Control, with Community Policies that govern what agentic AI tools can run, do, access, and reach.

Segment resources: - https://www.threatlocker.com/blog/the-principle-of-least-privilege-for-ai-agents - https://www.threatlocker.com/blog/applying-threatlocker-to-agentic-ai-tools - https://www.threatlocker.com/blog/why-the-five-eyes-alliance-sees-zero-trust-as-the-best-defense-against-agentic-ai-threats

This segment is sponsored by ThreatLocker. Visit https://securityweekly.com/threatlocker to learn more about them!

In the leadership and communications segment, 3 cybersecurity issues that should keep every CEO awake at night, You Don't Find Your Leadership Style. You Mentor Your Way Into It., Cybersecurity Starts With Communication – And We May Be Getting It Wrong, and more!

Show Notes: https://securityweekly.com/bsw-461

The Secret Word is Meow, Red Agent, GitHub, evoooo1bot, Hatman, DecryptAds, Copilot, Aaran Leyland, and More on the Security Weekly News.

Show Notes: https://securityweekly.com/swn-608