{ "@context": "https://schema.org", "@type": "BlogPosting", "@id": "https://mace.ai/blogs/mace-ai-founders-agent-security-thestreet/#article", "mainEntityOfPage": { "@type": "WebPage", "@id": "https://mace.ai/blogs/mace-ai-founders-agent-security-thestreet/" }, "url": "https://mace.ai/blogs/mace-ai-founders-agent-security-thestreet/", "headline": "Mace AI founders discuss AI agent security on TheStreet", "description": "Mace AI co-founders Greg Frank and Walton Comer join James Heckman on TheStreet Roundtable to discuss rogue AI agent swarms, permission boundaries, and the importance of independent adversarial testing for enterprise AI agents.", "datePublished": "2026-10-15", "inLanguage": "en-US", "articleSection": "In the media", "publisher": { "@type": "Organization", "name": "Mace AI", "url": "https://mace.ai/" }, "about": [ { "@type": "Thing", "name": "AI Agent Security" }, { "@type": "Thing", "name": "Agentic AI Security" }, { "@type": "Thing", "name": "Rogue AI Agent Swarms" }, { "@type": "Thing", "name": "AI Agent Permission Boundaries" }, { "@type": "Thing", "name": "Independent AI Security Testing" }, { "@type": "Thing", "name": "Adversarial AI Testing" }, { "@type": "Thing", "name": "Enterprise AI Risk Management" } ], "mentions": [ { "@type": "Person", "name": "Greg Frank", "jobTitle": "Co-founder and Chief Scientist", "worksFor": { "@type": "Organization", "name": "Mace AI", "url": "https://mace.ai/" } }, { "@type": "Person", "name": "Walton Comer", "jobTitle": "Co-founder and CEO", "worksFor": { "@type": "Organization", "name": "Mace AI", "url": "https://mace.ai/" } }, { "@type": "Person", "name": "James Heckman" }, { "@type": "Organization", "name": "TheStreet", "url": "https://www.thestreet.com/" }, { "@type": "Organization", "name": "OpenAI", "url": "https://openai.com/" }, { "@type": "Thing", "name": "AI Agent Permission Escalation", "description": "Security risks arising when AI agents are manipulated into performing actions beyond their authorized permissions." }, { "@type": "Thing", "name": "Independent AI Agent Testing", "description": "Evaluating deployed AI agents for security vulnerabilities, unauthorized actions, and permission-boundary violations." } ], "keywords": [ "Mace AI", "TheStreet Roundtable", "Greg Frank", "Walton Comer", "James Heckman", "AI Agent Security", "Rogue AI Agents", "AI Agent Swarms", "Agentic AI Security", "AI Agent Permissions", "AI Security Testing", "Adversarial AI Testing", "Enterprise AI Security", "AI Agent Risk Management", "AI Agent Governance" ], "isPartOf": { "@type": "Blog", "name": "Mace AI Blog", "url": "https://mace.ai/blogs/" } }
Skip to content
What we doSecurity & TrustResearchWhite PaperBenchmarksField GuideBlogs and ArticlesAbout
← All posts
In the media

Mace Ai founders discuss AI agent security on TheStreet

TheStreet Roundtable · October 15, 2026

Mace Ai co-founders Greg Frank and Walton Comer joined James Heckman on TheStreet Roundtable to discuss OpenAI’s Hugging Face breach by rogue agent swarms and what businesses can learn from agents acting beyond their assigned boundaries.

Mace Ai’s Chief Scientist, Greg Frank, explains why testing needs to examine what an agent does with the access it has been given, including whether it can be persuaded to exceed its permissions.

Walton Comer, CEO, shares his outlook on AI’s potential to help teams accomplish more. Their conversation explores the role of independent testing in helping businesses confidently put agents to work.

Greg Frank and Walton Comer on TheStreet Roundtable with James Heckman.
Read the full article on TheStreet →

Independent testing for the AI agents you deploy.

Get in touch →

More from the blog

Head-to-head
It reported '100% defended.' The agent was leaking the whole time.
Commentary
When safety training teaches deception
Research
Benchmark-robust isn't agent-safe.