MindFort AI

Last week, we evaled Grok 4.6 and 4.7 in our agent harness to test the models' ability to reason about, write, and execute real exploits against a target in a messy environment. Grok 4.6 scored incredibly well, becoming the best model we've ever tested at recon, pentesting, and reporting validated findings. Even more impressively, it did so at a $0.88 cost per finding, making it our 2nd most cost-efficient model for cyber use. See our updated findings here: mindfort.ai/research/nexbenc…

@mindfort can do this for you in one click. We also give you free credits to get started

Super excited to finally announce our raise! (We're hiring) We'll now be able to move much faster on these fronts: - increase our infra+capacity to serve a huge growth in demand - build out the team as we continue to develop new security agents for new verticals (cloud/network/remediations) - expand our applied research into offensive security LLMs Read more here on how we are thinking about the future of autonomous security agents and how MindFort is leading the push to a future where all companies are protected by teams of agents, inside and out. Huge thanks to @ycombinator @Soma_Capital and all our other early investors and angels for believing in us early. If you want to join us in building the future, send me a message, we are hiring across all roles in: Growth/Sales/Engineering/Research! mindfort.ai/blog/seed-anno… cc @AkulGupta30 @mindfort

>mom can we have chinese vulnerability discovery agent? >no honey, we have vulnerability discovery agent at home >@mindfort home grown in the USA

This market report on AI native AppSec companies (including @mindfort) points out where security is heading. Spoiler, it is heading to zero. Let me explain: As security agents rapidly mature and the cost of intelligence continues to drop per token, the average cost of discovery of a vulnerability will also drop. This is a double edged sword. - On one hand, it will collapse the vulnerability economy, making bounties and VDPs not worth the cost of maintaining them, instead deploying with low cost, highly effective agents. This will have an effect on the number of people working on security research. - On the other hand, it will also deflate the illicit market for vulnerabilities and reduce the number of zero days in circulation, and make discovery of exploitable vulnerabilities orders of magnitude more expensive for bad actors (who will also be using agents of their own). Companies will measure security in compute. Think: 100 kWs of compute, 100m tokens per day, 150 H100s powering your security program, etc. and can directly attribute better outcomes to increased compute spent on security. Spend becomes more efficient, finding vulns becomes cheaper, discovery happens faster, and overall security posture will increase. We are already measuring outcomes at MindFort this way. The cost of discovering a validated vulnerability used to be measured in 10s of millions of tokens. That is dropping to sub-million tokens on average using our latest agents and our upcoming model, MF-1. My most contrarian view on the topic is that the major AI labs will not win this market. A family of specialized cybersecurity models will win for a few reasons: - Bringing compute in house and inferencing for the lowest cost of vulnerabilities per compute unit - Ability to continuously train on your own security data, increasing that cost efficiency even further - Privacy and security: guaranteeing that model labs will not train on your data Now, while the cost of vulnerabilities will go to zero and AI agents take over continuous testing, this doesn't mean we will need less people in cyber: we will need more. We will need talented security researchers and engineers more than ever to run these programs at every company on earth. Report below: docs.google.com/document/d/1…

MindFort is delivering real results for engineering teams. Our agents help the team at Osto ship fast and keep their application secure.

MindFort goes beyond just discovering vulnerabilities. Introducing: CVSS 3.1 Composite Scoring Because our agents can deeply understand the context of web applications, they are not only great at finding and exploiting vulnerabilities, they are world class at vulnerability prioritization and scoring. MindFort takes into account many factors of an application relative to other vulnerabilities discovered and the application itself to make scoring and prioritization better than ever before. We even provide a coverage map of the OWASP Top 10, so you can understand your application's risk relative to industry standards. Teams report a significant reduction in triage time with MindFort Composite Risk Scoring, which is now generally available in all account tiers.

Soon, @mindfort agents will be able to test full standalone APIs. Great for engineering teams building and shipping developer APIs.

Jensen Huang just revealed something chilling about the future of AI and cybersecurity. On the latest episode of Joe Rogan, Jensen gave a stark warning about how today there are already millions of ai agents performing cyberattacks against friendly targets. And that the only way to stop them? Millions of our own agents fighting back. That is our mission at MindFort. To give you your own army of agents that are constantly attacking your targets, looking for a way in before the bad guys do. We don’t take this mission lightly, what we are building is critically dangerous, but we know that this is what is needed to keep the world safe. Super glad to hear Jensen, who goes on at lengeth, talk about both the danger and potential for AI in security. Excellent episode worth a watch. He is a brilliant person.

Had an absolute blast at @OffensiveAIcon yesterday! The speakers were increidble and it was great to catch up with old friends and new ones. Can’t wait for next year! 🌴 @mindfort @AkulGupta30

Honored to have @mindfort included in a16z’s article on the top AI security agent startups! a16z.com/next-gen-pentesting…