Security Weekly Podcast Network (Video)

In this week's interview, we kick off the conversation with how Sonrai's expertise in securing cloud identity permissions had the company well placed to address the explosion of AI agents and the clear risks they represented. On the surface, this looks like a cloud/hyperscaler permissions challenge, but it isn't that simple. As agents like Claude Code, Codex, and Hermes are connected to enterprise cloud agents, the risk spreads outside VPCs and onto endpoints.

Check out the episode to learn more about some of the most common risks Sandy finds and how Sonrai goes about addressing them.

This segment is sponsored by Sonrai Security. Visit https://securityweekly.com/sonrai to learn more about them!

Segment Resources

• AWS Bedrock agent permissions: what you need to lock down before you go live

Organizations looking to unlock the power of Enterprise AI Agents, and in a controlled and safe way at the speed of AI. Identity is at the heart of it. However, NHI Governance Is Not Enough for Enterprise AI Agents.

The identity industry has responded to the rise of AI agents the same way it responds to every new identity challenge: extend existing frameworks. Map agents to human owners. Enforce least privilege. Govern them like non-human identities.

It is a reasonable instinct. It is also insufficient in ways that matter enormously. Non-human identity security was built for a deterministic world - service accounts, API keys, bots. These identities do what they are configured to do. Their behavior is predictable enough that static governance models work. Enterprise AI agents are categorically different. Not in degree - in kind. They don't execute fixed instructions. They reason, plan, and adapt in response to context. Their scope shifts with every task. Their behavior at runtime can diverge significantly from anything true at provisioning time. Unlike any identity that came before them, they frequently change their intent, at a pace no governance model built for human movers or machine credentials was designed to handle.

Wrapping them in the same framework you use for a service account isn't wrong. It's just insufficient in precisely the places where risk accumulates.

• Download the SANS AI Security Maturity Model eBook

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

A self-driving car still has a license plate The accountability didn't change just because the driver did. The same has to be true for AI agents, but most environments can't trace an agent action back through the layers of delegation to the human who authorized it. Howard Ting, CEO of Opal Security, joins Security Weekly to discuss what the accountability model looks like when employees run swarms of agents, and what has to be in place before that accountability chain is tested.

• https://www.opal.dev/resource-center/identity-governance-report-2026-ai-access

This segment is sponsored by Opal Security. Visit https://securityweekly.com/opalidv to learn more about them!

Organizations have invested heavily in identity platforms, but many still struggle to maximize security, efficiency, and governance outcomes. As AI transforms both cyber defense and cyber threats, Identity Security is emerging as a critical foundation for securing human and non-human identities alike. In this discussion, we explore how AI is helping organizations reduce costs, improve operations, defend against AI-powered attacks, and address the governance challenges created by AI agents—highlighting the convergence of Identity Security, AI Security, and AI Governance.

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

Show Notes: https://securityweekly.com/esw-466

I talk to Gibb Witham, President of Hack The Box, about cyber readiness, hands-on security training, Hack The Box, and AI in cybersecurity.

Show Notes: https://securityweekly.com/swn-595

This week we have a technical segment based on the response to "Atomic Arch", an updated open-source tool to help you catch malicious packages. In the security news:

• Exploitarium
• A hot messy summer of vulnerabilities
• AI Squatting
• Linux LPE - no shortage of those
• Fingerprinting Favicons
• Windows 10 extended
• Can Clothes Make You Invisible to Facial Recognition?
• Fable and Mythos for All
• Do we care about Quantum?
• Execs have AI risk under control
• Biological warefare in Spyware
• The scripts in-scope for PCI
• We don't have privacy, but we may get age restrictions

Show Notes: https://securityweekly.com/psw-933

One of the biggest questions most executives ask is "Why does it still feel this hard when the talent is clearly there?" The answer, in almost every case, is not a people problem. It is an environment problem. And environment is something a leader can build.

Greg Hoffman, President at Ascension Performance Group, joins Business Security Weekly to discuss his new book, Performance Through People, a leadership parable that shows a practical operating model for building the conditions where people perform at their highest level. It is written as a story, but it is built as a framework. Greg will discuss the core pillars of this framework, including:

• Leadership First Mindset
• Operational Clarity
• Capability
• Empowerment
• Impact

Segment Resources: - https://a.co/d/053FuwYT - https://ascensionpg.com/articles/

In the leadership and communications segment, What the New Quantum Executive Orders Mean for Business Leaders, What I Learned About Burnout the Hard Way (and How to Actually Fix it, Mentorship Matters, and more!

Show Notes: https://securityweekly.com/bsw-454

AI Cocaine Recipes, Green Shirt Jailbreak, JLR Russia Hack, Scattered Spider, Cisco Root, Amazon Q Pwned – Aaran Leyland – SWN #594

Show Notes: https://securityweekly.com/swn-594

SquidBleed reveals another vuln that's been lurking for decades, but its real lesson is in managing an attack surface. Regardless of whatever programming language you use, removing code is one of the best security steps you can take, followed by changing default configs to turn off uncommon features and ancient protocols.

The Linux kernel's removal of strncpy is another example of managing attack surface by replacing a notoriously misused and ambiguous function with more specific versions that better match the developers intent. It was a six-year journey for the kernel, but one that should remove a class of vulns and, importantly, improve performance.

Then it's on to agents with a discussion of the newly released OWASP AISVS and yet another example of evaluating LLMs as code reviewers.

Agentic AI Has an Identity Problem

AI agents are already running inside enterprise environments, operating on credentials, API tokens, and cloud roles that most security teams have never inventoried. When an agent acts autonomously across production systems, the security question is no longer just what it can do but who it is and whether that identity is governed at all. Itamar Apelblat, Co-Founder and CEO of Token Security, discusses why identity is the right lens for understanding agentic AI risk and what practical steps security teams can take now.

Segment Resources:

• https://www.token.security/product
• https://www.token.security/lp/ai-agent-identity-security-buyers-guide-ebook
• https://www.token.security/enzo
• https://www.token.security/ai-agent-calculator

This segment is sponsored by Token Security. To lean more, visit https://securityweekly.com/tokenidv

Blended Identities and the challenge of IAM for AI

AI agents aren't quite human and aren't traditional machines. So how do you secure workflows that involve humans using AI to access sensitive data, and do it at machine speed and scale? David breaks down the challenges and discusses actual implementations of IAM for AI to explain how to solve them.

Segment Resources:

• https://aembit.io/case-study/a-300b-investment-firm-secures-claude-access-with-aembit/
• https://aembit.io/blog/aembit-now-secures-microsoft-copilot-studio-agents/
• https://www.youtube.com/watch?v=cSInzRUXvNc

This segment is sponsored by Aembit. Get the cloud security alliance survey on AI Identities at https://securityweekly.com/aembitidv

Show Notes: https://securityweekly.com/asw-389

Adriel joins us for a discussion on the state of penetration testing, why it hasn't done much to help security teams over the last 20 years, and why AI won't save it.

Segment Resources:

• https://hbr.org/2026/04/boards-are-falling-short-on-cybersecurity
• https://www.scworld.com/perspective/how-to-build-a-breach-ready-security-posture-without-the-enterprise-price-tag
• https://netragard.com/blog/what-is-penetration-testing/

The titular essay: https://newsletter.pragmaticengineer.com/p/why-is-meta-destroying-its-engineering

A very interesting analysis of what's going on inside big tech companies as they try to dogfood their own AI hype and tokenmaxx themselves into oblivion. There have been a LOT of stories on this, but this is the most comprehensive and enlightening. A few more are linked below.

This is relevant to security, because heavier AI use appears to be linked to a much higher occurrence of availability and security issues.

• 'Tell Him He's a Piece of Shit': Meta's New AI Unit Is a Total Mess
• The Newest Instagram "Exploit" is the Goofiest I've Seen
• Meta CTO Andrew Bosworth Admits the Company's AI Reorg Was 'Atrocious'
• Meta's months-old AI unit is a soul-crushing gulag, say the engineers stuck inside it

Finally, in the enterprise security news,

1. an AI vibe check
2. An AI SOC vendor shuts down
3. Cybersecurity vendor layoffs
4. funding & acquisitions
5. cascading breaches
6. digital estate management
7. criminals don't trust AI either
8. some devs won't code without AI, even if you pay them to
9. Midjourney is now a healthcare company?

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

Show Notes: https://securityweekly.com/esw-465

AI Brain Harvest, Fortibleed, Win 10, Blacksite, Windchill, Cisco, BB-8 Sidewalk Bots, Josh Marpet, and More on this episode of the Security Weekly News.

Show Notes: https://securityweekly.com/swn-593

First up is Sandy Bird from Sonrai discussing how to protect our cloud infrastructure!

This segment is sponsored by Sonrai Security. Visit https://securityweekly.com/sonrai to learn more about them!

Next up in the security news:

• Help, I am Fortibleeding
• Cisco SD-WAN needs help
• The secret life of probe requests
• Help, I am Squidbleeding
• XSS to RCE and why CVSS isn't the full picture
• TVs spy on you
• Foundational security practices
• Cybersecurity costs money
• Happy "Its too late to update your KEK key" day
• You don't have security flaws if no one can report them
• Rickrolling FIFA
• Domain takeovers
• End of life, out of luck
• The key to Encryption...

Show Notes: https://securityweekly.com/psw-932

The 2026 Verizon DBIR has arrived and the results are in... Even with a substantial increase in Exploitation of Vulnerabilities, All Credential Abuse is still the top initial access vector for breaches, which means the human is still the weakest link. Why haven't security awareness training and phishing campaigns worked?

Robert Siciliano, Architect of of The Strategic Human Firewall™ at ProtectNow, joins Business Security Weekly to explore why humans, not hackers, are the ultimate deciding factor in organizational security. The industry needs to shift from security awareness to security appreciation. Robert will discuss:

• How you can build a culture that actually protects your people, your data, and your operations in an era of AI deception.
• Why most companies are still performing 'Security Theater'—checking boxes and hoping for the best—instead of driving genuine behavior change.
• How Trust and Denial quietly fuel most disasters, why interactive training is the only way to make the lessons stick, and how leaders can scale this entire framework without needing a Hollywood budget.

Segment Resources:

https://protectnowllc.com/ai-cyber-security-keynote-speaker/

In the leadership and communications segment, Should CEOs Be Held Personally Accountable for Cyber Attacks?, Placing communication at the center of every leadership transition, AI isn't solving cybersecurity workforce woes, and more!

Show Notes: https://securityweekly.com/bsw-453