A leaked credential shows up in a criminal marketplace, or a vulnerability gets a disclosure advisory, and either one can be weaponized against a real target before most security teams have triaged the alert. Attackers are combining that kind of intelligence with AI-assisted exploitation to accelerate the path from exposure to breach faster than most security programs are built to react.
Intelligence is still the earliest signal defenders get, and a leaked credential turning up in a feed is proof of how useful that signal has become. The problem sits one step later, in what happens after the signal arrives.
The Queue Where Risk Accumulates
In most organizations, a high-value indicator waits in a queue instead of getting acted on right away, until someone with the offensive skill to test it actually has the time to determine whether it’s exploitable in that specific environment, on that specific day. That queue, more than any shortage of intelligence, is where exposure builds up.
This shows up on both sides of the industry. Security teams describe it as a backlog problem. Speaking with the product teams of Recorded Future, the world’s largest threat intelligence company, I’ve heard the same pattern from their side: the volume of relevant threat data outpaces most teams’ capacity to test each item against a live environment, and validating at that scale is limited by time and specialized offensive skill, not by a lack of data.
From Probability to Proof
That’s part of why threat-led penetration testing, TLPT, moves beyond a compliance requirement in a handful of regulated sectors and into a broader operating model. TLPT starts from what current intelligence says is actually happening: a specific leaked credential, a specific disclosed vulnerability; and tests for that directly instead of working through a static backlog on a fixed calendar.
For a leaked credential specifically, TLPT is built to return evidence: this exact credential is or isn’t exploitable in this exact environment right now. That’s where a lot of security teams say they want to spend their limited testing capacity.
What This Looks Like in Practice
Pentera’s collaboration with Recorded Future is one example of this shift taking shape, and it’s the one I know best. The integration is built so that a threat signal, whether it originates from Recorded Future, from Pentera’s own platform, or from another intelligence source, can trigger an automated validation run against the organization’s real attack surface. The first capability built on this connects Recorded Future’s leaked-credential intelligence to automated testing of an organization’s external attack surface, confirming which exposed credentials can be used by an attacker to exploit a certain environment, rather than flagging all of them as equally urgent.
One of the customers involved in early testing described the shift in plain terms. Joseph Gothelf, Vice President of Cybersecurity at Wyndham Hotels & Resorts, said: “The convergence of threat intelligence and security validation is one of the most important shifts in our security program. Knowing what’s coming is only half the answer. Being able to test against it in our own environment, at speed, is what builds real resilience in the AI era.”
Recorded Future’s feed surfaces a leaked credential the same way it would for any customer running it. Whether that specific credential still works against a specific environment, regardless of which vendors are involved in surfacing or testing it, is what most security programs still can’t answer quickly. That’s the gap I’d put security budget against before another intelligence feed: not knowing more, but proving what’s already known.
To learn more, join the “Threat Intel’ Just Got Teeth, TLPT Goes Live” webinar on September 29.
Note: This article has been expertly written and contributed by Doron Naim, VP Strategic Alliances, Pentera.
