AI Solutions Resources About Talk to us Log in Try Platform Try Platform Platform overview Pricing Handwriting recognition Table extraction Data transformations Straight-through processing Data matching Financial Services Insurance Logistics View all Dynamics 365 Salesforce Xero View integration pathways Blog Customer stories Expert insights White papers Affinda Academy API reference Product guides Product updates Security Trust Centre About Affinda Pathfindr Careers Media & awards Customer stories Platform overview Pricing Handwriting recognition Table extraction Data transformations Straight-through processing Data matching Financial Services Insurance Logistics View all Dynamics 365 Salesforce Xero View integration pathways Blog Customer stories Expert insights White papers Affinda Academy API reference Product guides Product updates Security Trust Centre About Affinda Pathfindr Careers Media & awards Customer stories Log in /* Hide-on-scroll: past 100px, hide when scrolling down, reveal on up. */ (function () { let lastY = window.scrollY; let ticking = false; function onScroll() { if (ticking) return; ticking = true; requestAnimationFrame(function () { const y = window.scrollY; document.querySelectorAll('af-navbar').forEach(function (nb) { if (y > 100 && y > lastY) nb.classList.add('nav-hidden'); else nb.classList.remove('nav-hidden'); }); lastY = y; ticking = false; }); } window.addEventListener('scroll', onScroll, { passive: true }); })(); /* Tag external dropdown links to open in a new tab. af-nav-item renders its anchor internally and exposes no `target` prop, so we post-process the slotted nav-items after Stencil hydrates. A MutationObserver scoped to the navbar covers the initial render and any later re-renders (mobile menu open/close, etc.). */ (function () { function tag(root) { root.querySelectorAll('af-nav-item a[href^="http"]').forEach(function (a) { if (a.getAttribute('target') !== '_blank') { a.setAttribute('target', '_blank'); a.setAttribute('rel', 'noopener noreferrer'); } }); } document.querySelectorAll('af-navbar').forEach(function (nb) { tag(nb); new MutationObserver(function () { tag(nb); }).observe(nb, { childList: true, subtree: true, }); }); })(); ← All expert insights Tech AI When my AI agent hacked my gym, Mythos stopped feeling theoretical I gave an AI agent permission to book gym classes. It found authorization vulnerabilities in a major software provider. That experience made Anthropic's Mythos announcement feel very real.
This was not a grand research project. It was a practical little automation. The classes fill up fast, I got tired of playing refresh roulette, and I figured an agent running on Opus 4.6 could handle the annoying part for me.
In the course of trying to book classes, the bot discovered that the gym software provider exposed a GraphQL API with authorization flaws. Not tiny edge-case flaws, either. It could book classes months outside the intended booking window, before they were supposed to be available. Worse, it could cancel other members's reservations and bump them off the waitlist.
What made the whole thing more surreal was the tone. The bot was not malicious. It was helpful. After finding the issue, it drafted a responsible disclosure email to support, explained the vulnerability, suggested fixes, and even compared the broken mutations with the ones that correctly enforced authorization. I had to tell it to write that email, which is worth noting. But the whole experience gave me a very visceral feeling that I think a lot of people still do not quite have yet: if you give an AI agent permission to go do the thing, it will often discover paths you did not explicitly ask it to look for.
If you only read the headlines, it sounds like another "new model is better at benchmarks" story. I do not think that is what is interesting here. What is interesting is the shape of the capability. Anthropic says Mythos is around 40 percent above Opus 4.6 on the benchmarks they care about, and they are not releasing it publicly because they think it is too dangerous. Instead, they put it behind a limited program with a small group of partners to defensively scan for vulnerabilities.
This article was aggregated automatically by CyberWire Daily's newsfeed engine. Original reporting: web.archive.org.
