Shadow AI & Sandbox Escapes: Why You Need an Agentic Control Plane?

Shadow AI & Sandbox Escapes: Why You Need an Agentic Control Plane?

When Claude Cowork hits a roadblock, it doesn't give up, it writes a custom Python script and downloads an untrusted NPM package just to bypass its restrictions and finish its goal. Are your security tools close enough to stop it? In this episode, Ashish sits down with Michael Leland, VP, Field CTO at Island, to discuss the critical need for an Agentic Control Plane. Michael breaks down why traditional security silos (EDR, DLP, CASB) fail to provide visibility when autonomous AI agents execute tasks outside the network, and why the browser is the ultimate line of defense for monitoring user intent. We explore the massive reality of Shadow AI and the hidden danger of well-intentioned employees accidentally hooking up sensitive data to public LLMs. Finally, Michael shares practical strategies for solving token waste through "model fit steering" and managing the complex "two-hop problem" when an agent calls another agent.


Guest Socials -⁠⁠ ⁠⁠⁠⁠⁠⁠⁠⁠⁠Michael's Linkedin

Podcast Twitter - ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠@CloudSecPod⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠

If you want to watch videos of this LIVE STREAMED episode and past episodes - Check out our other Cloud Security Social Channels:

-⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Cloud Security Podcast- Youtube⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠

- ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Cloud Security Newsletter ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠

If you are interested in AI Security, you can check out our sister podcast -⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ AI Security Podcast⁠


Questions asked:

(00:00) Introduction to the Agentic Control Plane(01:50) Michael Leland’s Background (Cabletron, Nitro Security, SentinelOne)(03:20) Why Island Evolved from the Browser to the Desktop for AI(05:50) The Failure of Traditional Siloed Security (EDR, DLP, CASB)(07:20) Goal-Oriented AI: How Claude Cowork Downloads Untrusted NPM Packages(08:30) Model Fit Steering: Routing Users to the Right LLM for the Right Price(10:30) The Threat of Malicious AI Skills and Plugins(11:30) The Well-Intentioned Insider Threat (The Next Cambridge Analytica)(13:00) Uncovering Shadow AI: From 8 Tools to 243(15:10) Token Brokering at the MCP Gateway(16:40) Protecting Non-Human Identities (NHI)(18:20) Solving the "Two-Hop" Problem (Agent-to-Agent Communication)(21:00) Fixing Hallucinations with Corporate RAGs(25:40) Calculating AI ROI Beyond "Token Maxing"(28:40) The "You Laugh, You Lose" Cybersecurity Joke Challenge

Det här avsnittet är hämtat från ett öppet RSS-flöde och publiceras inte av Podme. Det kan innehålla reklam.

Avsnitt(365)

Pentesting AWS Security Agent

Pentesting AWS Security Agent

When cloud providers introduce autonomous AI agents to conduct penetration testing on customer environments, who is responsible when the AI breaks out of scope?In this episode of the Cloud Security Po...

5 Nov 38min

Using LLMs for Enterprise Threat Modeling

Using LLMs for Enterprise Threat Modeling

Creating an application threat model was a resource-intensive process that took weeks or months, resulting in static documents that were abandoned 80% of the time after their initial creation. However...

27 Okt 35min

Mythos Hype vs. Reality: Why You Need an Agentic Harness

Mythos Hype vs. Reality: Why You Need an Agentic Harness

Is the fear surrounding Anthropic’s Mythos model justified? In this episode, Ashish sits down with Johnny Hand, Global CISO at Trend AI and host of The AI Security Brief. Johnny breaks down why turnin...

21 Okt 47min

How Block Solved DNS Vulnerabilities

How Block Solved DNS Vulnerabilities

When companies rapidly migrate to the cloud without strict registrar policies, users often spin up unsanctioned subdomains pointed at third-party services like Vercel or Hugging Face. When those servi...

13 Okt 55min

Why AI Security is Actually a Data Security Problem

Why AI Security is Actually a Data Security Problem

When an AI coding agent hallucinates a software package name, attackers can simply create malicious packages with that exact name, tricking the AI into downloading malware directly into your environme...

8 Okt 40min

Securing Millions of AI-built Apps: How Lovable Defends Against Insider Threats

Securing Millions of AI-built Apps: How Lovable Defends Against Insider Threats

Empowering non-technical users to build production-grade applications requires a proactive, secure-by-design architecture. In this episode, Ashish sits down with Marcus Hallberg and Samuel Kelemen, se...

11 Sep 1h 10min

Why AI Won't Replace Your SOC: Federated Data & APEX Framework

Why AI Won't Replace Your SOC: Federated Data & APEX Framework

Hash values, IP addresses, and domains are virtually useless as primary detection mechanisms in the era of AI-driven polymorphic malware and short-lived phishing kits. If your detection strategy is st...

1 Sep 37min

The "Hunt First" AI Security Strategy

The "Hunt First" AI Security Strategy

Did you know that 82% of all intrusions don't involve any sort of malware, and AI-augmented attacks are up 89% year over year? If your security operations team is solely focused on reacting to known-b...

25 Aug 46min

Populärt inom Teknik

uppgang-och-fall
natets-morka-sida
bilar-med-sladd
elbilsveckan
market-makers
rss-veckans-ai
skogsforum-podcast
rss-ai-med-jonas-benjamin
rss-snacka-om-ai
rss-laddstationen-med-elbilen-i-sverige
rss-elektrikerpodden
vi-bilagares-podcast
rss-technokratin
bosse-bildoktorn-och-hasse-p
rss-it-sakerhetspodden
bli-saker-podden
hej-bruksbil
rss-uppgang-och-fall
rss-kack-tech-podcast
ai-sweden-podcast