{"id":243482,"date":"2026-07-15T11:38:00","date_gmt":"2026-07-15T15:38:00","guid":{"rendered":"https:\/\/testing.news-you-need.com\/index.php\/2026\/07\/15\/why-ai-harnesses-matter-more-than-frontier-llms-for-cybersecurity\/"},"modified":"2026-07-15T11:40:08","modified_gmt":"2026-07-15T15:40:08","slug":"why-ai-harnesses-matter-more-than-frontier-llms-for-cybersecurity","status":"publish","type":"post","link":"https:\/\/testing.news-you-need.com\/index.php\/2026\/07\/15\/why-ai-harnesses-matter-more-than-frontier-llms-for-cybersecurity\/","title":{"rendered":"Why AI ‘harnesses’ matter more than frontier LLMs for cybersecurity"},"content":{"rendered":"

Why AI ‘harnesses’ matter more than frontier LLMs for cybersecurity<\/a><\/p>\n

https:\/\/cyberscoop.com\/ai-cybersecurity-harness-autonomous-hacking\/<\/a><\/p>\n

Publish Date: 2026-07-15 11:38:00<\/a><\/p>\n

Source Domain: cyberscoop.com<\/a><\/p>\n

Author: <\/a><\/p>\n

Using an unordered list, summarize the following article with between 4 and 8 key points. <\/p>\n

As AI-enabled hacking becomes a bigger threat for cybersecurity and national security, public attention has focused on mainly a few leading frontier AI companies developing more powerful large language models.<\/p>\n

These models, and the billions of dollars behind them matter, but they\u2019re only part of a larger shift. Enterprises are now building their own technology platforms that take these general-purpose LLMs and turn them into bespoke cybersecurity tools.<\/p>\n

Industry professionals refer to these tools as a \u201charness.\u201d\u00a0They control the model\u2019s behavior, limit its risks, and connect it\u00a0to internal IT systems and networks so it can work reliably at scale.<\/p>\n

New research from Cato Networks shared exclusively with CyberScoop shows how much power can come from a harness. It paired OpenAI\u2019s ChatGPT 5.5 and GPT 5.5-Cyber models with its own tool and tested the abilities of the agent to hack into a victim network with as little human direction as possible.<\/p>\n

Across six different scenarios, the pairing achieved complete end-to-end attack chains, including domain administrator privileges and Active Directory access, sometimes in as little as 40 minutes.<\/p>\n

\u201cWhat was most surprising is that first we saw that it was capable of doing accelerated reasoning and attack, and interacting and doing all this by itself, like doing all of the stages of the attacks,\u201d said Guy Weisel, a tech evangelist at Cato Networks and one of the authors behind the research.<\/p>\n

Critically, the most successful scenarios happened when the model was given appropriate operational context from the technical harness developed by Cato Networks.\u201cIt does support that it\u2019s not just about the frontier model,\u201d said Weisel. \u201cWe found that [our harness] really helps the reasoning\u201d of the LLM.<\/p>\n

An illustration of an agentic AI attack chain and lateral movement within victim networks. (Source: Cato Networks)<\/p>\n

The agent was given some \u2013 but not abundant \u2013 resources to complete its tasks, including an external Kali Linux attack host, the simulated target\u2019s public IP address and a set of low-level domain credentials acquired through phishing.<\/p>\n

It was not provided with any other details, and had to probe further for key information, such as further knowledge of the server type (Microsoft Exchange), the target\u2019s operating system, version, build number, internal network topology, access to higher privilege accounts and other critical assets, nor was agent given any predetermined attack paths.<\/p>\n

The Cato Networks research uses OpenAI models, but only as an example. Weisel said he believes other models would likely achieve similar results. In any event, if current trends hold, the kind of capabilities provided by LLMs like GPT 5.5 are likely to be open-source within a year.<\/p>\n

Cato Networks is far from alone. Most enterprises have their own AI harnesses, and\u00a0 executives tell CyberScoop they are playing an increasing role in more effectively steering the frontier model workflows.<\/p>\n

While AI tools can struggle to duplicate human workflows in other areas, LLMs have long shown potential in cybersecurity and coding, improving greatly over the past few years. The Trump administration has set up a new federal clearinghouse for exchanging information between the public and private sectors on AI-discovered vulnerabilities, while European groups are setting up their own organizations to coordinate globally on AI cyber threats.<\/p>\n

Eric Doerr, chief product officer at Tenable, told CyberScoop a harness used in the company called \u201cHexa\u201d\u00a0 offers a defensive advantage:\u00a0 it can work with\u00a0different commercial LLMs while delivering consistent\u00a0 results.<\/p>\n

\u201cOne of the first things we do when we get a [new] model is say \u2018Well, let\u2019s run it through Hexa and see what we learn,\u2019\u201d said Doerr. \u201cWe have a whole bunch of benchmarks. Is it the same, is it better? Where is it better? Where is it worse?\u201d<\/p>\n

Hexa is meant to ensure that whichever model or models become dominant, Tenable will be able to integrate it into their tech stack and protect their most sensitive assets from unintended behaviors. That frees up the LLM to do what it does best: find vulnerable code and establish attacker pathways for exploiting them.<\/p>\n

\u201cFor years, it has been true that there are way more potential issues that a company has to deal with: code vulnerabilities, things that are unpatched, misconfigurations,\u201d said Doerr. \u201cThere\u2019s way more than you can actually remediate, and you really need to understand the difference between what\u2019s a theoretical problem and a real problem.\u201d<\/p>\n

Dan Rapp, chief AI and data officer at Proofpoint, said their harness, \u201cSatori,\u201d has become a critical tool for keeping their agentic AI on track while giving humans the ability to step in when things go awry.<\/p>\n

\u201cI think what you\u2019re seeing in the foundation of frontier models is you have raw intelligence, raw reasoning power, but to get these systems to perform the way you want to, both context engineering \u2013 the content provided ensuring that its accurate and relevant \u2013 and the harness engineering are essential to actually get the systems to perform well,\u201d Rapp told CyberScoop.<\/p>\n

That was a common theme in interviews with companies. While frontier models come and go, or are overtaken by international competitors, there will always be the need for the model to operate with data and context that often only the organization can provide.\u00a0\u00a0<\/p>\n

It suggests that while policymakers and cybersecurity experts have focused on the spread of newer and more powerful frontier models, industry \u2013 and likely soon the cybercriminal underground \u2014 has quickly developed the kind of technical infrastructure that is becoming far more important to AI cyber defensive and offensive tasks.<\/p>\n

\u201cWe\u2019ve had to bootstrap quite a few of these systems from first principles, and what it always boils down to is how effective you are with the tool calling\u2026 bringing in data, enriching the context,\u201d said John Hopper, vice president of product engineering at SpecterOps.<\/p>\n

\t\t\tWritten by Derek B. Johnson
\n\t\t\tDerek B. Johnson is a reporter at CyberScoop, where his beat includes cybersecurity, elections and the federal government. Prior to that, he has provided award-winning coverage of cybersecurity news across the public and private sectors for various publications since 2017. Derek has a bachelor\u2019s degree in print journalism from Hofstra University in New York and a master\u2019s degree in public policy from George Mason University in Virginia.\t\t<\/p>\n

<\/p>\n","protected":false},"excerpt":{"rendered":"

Why AI ‘harnesses’ matter more than frontier LLMs for cybersecurity https:\/\/cyberscoop.com\/ai-cybersecurity-harness-autonomous-hacking\/ Publish Date: 2026-07-15 11:38:00…<\/p>\n","protected":false},"author":1,"featured_media":243483,"comment_status":"closed","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"fifu_image_url":"https:\/\/cyberscoop.com\/wp-content\/uploads\/sites\/3\/2026\/07\/AI-harness.jpeg","fifu_image_alt":"","footnotes":""},"categories":[15],"tags":[26,24,17,25],"class_list":["post-243482","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-cybersecurity","tag-ai","tag-cybersecurity","tag-llm","tag-phishing"],"_links":{"self":[{"href":"https:\/\/testing.news-you-need.com\/index.php\/wp-json\/wp\/v2\/posts\/243482"}],"collection":[{"href":"https:\/\/testing.news-you-need.com\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/testing.news-you-need.com\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/testing.news-you-need.com\/index.php\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/testing.news-you-need.com\/index.php\/wp-json\/wp\/v2\/comments?post=243482"}],"version-history":[{"count":1,"href":"https:\/\/testing.news-you-need.com\/index.php\/wp-json\/wp\/v2\/posts\/243482\/revisions"}],"predecessor-version":[{"id":243484,"href":"https:\/\/testing.news-you-need.com\/index.php\/wp-json\/wp\/v2\/posts\/243482\/revisions\/243484"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/testing.news-you-need.com\/index.php\/wp-json\/wp\/v2\/media\/243483"}],"wp:attachment":[{"href":"https:\/\/testing.news-you-need.com\/index.php\/wp-json\/wp\/v2\/media?parent=243482"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/testing.news-you-need.com\/index.php\/wp-json\/wp\/v2\/categories?post=243482"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/testing.news-you-need.com\/index.php\/wp-json\/wp\/v2\/tags?post=243482"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}