[{"data":1,"prerenderedAt":3420},["ShallowReactive",2],{"application-flags":3,"navbar":7,"always-visible-banner":36,"navbar-about-highlight":108,"navbar-resource-highlight":182,"blog/how-new-ai-agents-will-transform-credential-stuffing-attacks":226},[4],{"name":5,"enabled":6},"maintenanceMode",false,[8],{"createdDate":9,"id":10,"name":11,"modelId":12,"published":13,"query":14,"data":15,"variations":20,"lastUpdated":21,"firstPublished":22,"testRatio":23,"createdBy":24,"lastUpdatedBy":25,"folders":26,"meta":27,"rev":35},1742208588866,"1c7a4e423bf54ac1a328bb4063459ef2","Banner","1c6207a5f24948ab82d4a0b17f251193","published",[],{"type":16,"url":17,"text":18,"link":19},"web-banner","https://pushsecurity.com/resources/browser-attacks-report","Get our latest report analyzing browser attack techniques in 2026",{},{},1774258294825,1742208637545,1,"CydmZnOWU1XuAaLhEDCoYNM4Z8W2","jKjF9r5jcvXU8tzZEfFQm31Iyvr2",[],{"kind":28,"lastPreviewUrl":29,"breakpoints":30,"hasAutosaves":34},"data","",{"xsmall":31,"small":32,"medium":33},320,640,768,true,"ga0kpxjhh76",{"createdDate":37,"id":38,"name":39,"modelId":40,"published":13,"stageModifiedSincePublish":6,"query":41,"data":42,"variations":97,"lastUpdated":98,"firstPublished":99,"testRatio":23,"createdBy":100,"lastUpdatedBy":101,"folders":102,"meta":103,"rev":107},1774965361051,"fd266d0172cc47429be7ad10f48c99ad","always visible banner","0678d178ec8b41efb8a23c09dba7874d",[],{"ctaText":43,"text":44,"url":29,"blocks":45,"state":93},"ewrererw","testrfesssssssssss",[46,73,81],{"@type":47,"@version":48,"id":49,"component":50,"responsiveStyles":63},"@builder.io/sdk:Element",2,"builder-ca12c06a52de41d7b8743da53118cd38",{"name":51,"tag":51,"options":52,"isRSC":62},"TopBannerContent",{"text":53,"ctaText":54,"url":55,"mainText":56,"cta":59},"New Webinar Series: Join John Hammond, Troy Hunt, and Matt Johansen for the State of Browser Attacks","Save Your Spot","https://pushsecurity.com/webinar/state-of-browser-security",{"content":57,"fontSize":58},"\u003Cp>Meet Push's browser security experts at BlackHat 2026.\u003C/p>","text-base",{"content":60,"fontSize":58,"url":61},"\u003Cp>Book a meeting →\u003C/p>","https://pushsecurity.com/events/blackhat-2026-meeting",null,{"large":64},{"display":65,"flexDirection":66,"position":67,"flexShrink":68,"boxSizing":69,"marginTop":70,"marginBottom":70,"fontSize":71,"fontWeight":72},"flex","column","relative","0","border-box",".56rem","1.125rem","700",{"@type":47,"@version":48,"id":74,"component":75,"responsiveStyles":79},"builder-a2e1f4b9f30b464bb814d7f5de5b0aa7",{"name":76,"options":77,"isRSC":62},"Custom Code",{"code":78,"scriptsClientOnly":6},"\u003Cstyle>\n  .top-banner.bg-web-orange{background:rgb(114, 79, 255);}\n\u003C/style>\n",{"large":80},{"display":65,"flexDirection":66,"position":67,"flexShrink":68,"boxSizing":69},{"id":82,"@type":47,"tagName":83,"properties":84,"responsiveStyles":88},"builder-pixel-o503s4fpvk","img",{"src":85,"aria-hidden":86,"alt":29,"role":87,"width":68,"height":68},"https://cdn.builder.io/api/v1/pixel?apiKey=f3a1111ff5be48cdbb123cd9f5795a05","true","presentation",{"large":89},{"height":68,"width":68,"display":90,"opacity":68,"overflow":91,"pointerEvents":92},"block","hidden","none",{"deviceSize":94,"location":95},"large",{"path":29,"query":96},{},{},1783541846936,1774968080803,"ST0tXQM8slWpFrmioqKHmENB2qe2","kYgMv6WsbvfmlOUYqR2SFwGzw6e2",[],{"kind":104,"lastPreviewUrl":105,"hasLinks":6,"breakpoints":106,"hasErrors":6,"hasAutosaves":6},"component","https://pushsecurity.com/?builder.space=f3a1111ff5be48cdbb123cd9f5795a05&builder.user.permissions=read%2Ccreate%2Cpublish%2CeditDesigns%2CeditLayouts%2CeditLayers%2CeditContentPriority%2CeditFolders%2CcreateProjects%2CsendPullRequests&builder.user.role.name=Designer&builder.user.role.id=creator&builder.cachebust=true&builder.preview=always-visible-banner&builder.noCache=true&builder.allowTextEdit=true&__builder_editing__=true&builder.overrides.always-visible-banner=fd266d0172cc47429be7ad10f48c99ad&builder.overrides.fd266d0172cc47429be7ad10f48c99ad=fd266d0172cc47429be7ad10f48c99ad&builder.options.locale=Default",{"xsmall":31,"small":32,"medium":33},"kyujw1aptq",[109,145],{"createdDate":110,"id":111,"name":112,"modelId":113,"published":13,"stageModifiedSincePublish":6,"query":114,"data":115,"variations":138,"lastUpdated":139,"firstPublished":140,"testRatio":23,"createdBy":24,"lastUpdatedBy":24,"folders":141,"meta":142,"rev":144},1776247359804,"9136a8f18b3b4a6ba29b8653a99372b1","testimonial-inductive-automation","20d9eaa352304613b3d1a794b400703d",[],{"link":116,"type":117,"testimonialLink":118,"testimonial":119},{},"testimonial","/customer-stories/inductive-automation",{"@type":120,"id":121,"model":117,"value":122},"@builder.io/core:Reference","f028f2b685bb47cd8bf9e82a26dd5a79",{"query":123,"folders":124,"createdDate":125,"id":121,"name":126,"modelId":127,"published":13,"data":128,"variations":132,"lastUpdated":133,"firstPublished":134,"testRatio":23,"createdBy":100,"lastUpdatedBy":100,"meta":135,"rev":137},[],[],1735823466309,"We found Push to be more accurate when compared to competitors and the browser agent offered features that others couldn’t match.","42035571a56940ac98bff4544aa79aa5",{"author":129,"jobTitle":130,"quote":126,"image":131},"Jason Waits","\u003Cp>CISO at Inductive Automation\u003C/p>","https://cdn.builder.io/api/v1/image/assets%2Ff3a1111ff5be48cdbb123cd9f5795a05%2Ff04c0c0689ce4a89ac0f0708d78c0a07",{},1735910703862,1735823501152,{"kind":28,"lastPreviewUrl":29,"breakpoints":136,"hasAutosaves":34},{"small":32,"medium":33},"e8bsg4qrgtu",{},1776247404986,1776247404973,[],{"breakpoints":143,"kind":28,"lastPreviewUrl":29,"hasAutosaves":6},{"xsmall":31,"small":32,"medium":33},"g9lqcuxphw7",{"createdDate":146,"id":147,"name":148,"modelId":113,"published":13,"meta":149,"stageModifiedSincePublish":6,"query":151,"data":152,"variations":178,"lastUpdated":179,"firstPublished":180,"testRatio":23,"createdBy":24,"lastUpdatedBy":24,"folders":181,"rev":144},1776255761419,"05a9322735fc427db12e2740e4302300","Report: 2026 Browser Attack Techniques",{"breakpoints":150,"kind":28,"lastPreviewUrl":29,"hasAutosaves":6},{"xsmall":31,"small":32,"medium":33},[],{"testimonial":153,"link":172,"type":175,"title":148,"description":176,"image":177},{"@type":120,"id":154,"model":117,"value":155},"192acbb1f9ca4cac918c0ec435a8bae3",{"query":156,"folders":157,"createdDate":158,"id":154,"name":159,"modelId":127,"published":13,"data":160,"variations":166,"lastUpdated":167,"firstPublished":168,"testRatio":23,"createdBy":100,"lastUpdatedBy":24,"meta":169,"rev":171},[],[],1728981467463,"Push does for identity what CrowdStrike did for the endpoint",{"video":161,"jobTitle":162,"author":163,"qoute":29,"quote":164,"image":165},"https://cdn.builder.io/o/assets%2Ff3a1111ff5be48cdbb123cd9f5795a05%2F8b30e8ca50064058bbaef0f3c6164575%2Fcompressed?apiKey=f3a1111ff5be48cdbb123cd9f5795a05&token=8b30e8ca50064058bbaef0f3c6164575&alt=media&optimized=true","\u003Cp>Deputy CISO at Microsoft\u003C/p>\u003Cp>Former LinkedIn, Slack, Palantir\u003C/p>","Geoff Belknap","Push does for identity what CrowdStrike did for the endpoint.","https://cdn.builder.io/api/v1/image/assets%2Ff3a1111ff5be48cdbb123cd9f5795a05%2F748f0ad0a5064a00a13f4721fcc8dea1",{},1742902158597,1728981782923,{"kind":28,"lastPreviewUrl":29,"breakpoints":170,"hasAutosaves":34},{"small":32,"medium":33},"ulhc0im1hfo",{"text":173,"url":174},"Download now","/resources/browser-attacks-report","resource","Learn about the latest techniques being used in the wild.","https://cdn.builder.io/api/v1/image/assets%2Ff3a1111ff5be48cdbb123cd9f5795a05%2F7b4a5ebf81d64e8c9d7fc35f6c96c4a9",{},1776255810913,1776255810900,[],[183,205],{"createdDate":184,"id":185,"name":148,"modelId":186,"published":13,"meta":187,"stageModifiedSincePublish":6,"query":189,"data":190,"variations":200,"lastUpdated":201,"firstPublished":202,"testRatio":23,"createdBy":24,"lastUpdatedBy":24,"folders":203,"rev":204},1776256900280,"1f429607996e4e5fae8fe3f9b9610e55","4829faa81e7c4ee8bd2d000e160e8d3c",{"breakpoints":188,"kind":28,"lastPreviewUrl":29,"hasAutosaves":6},{"xsmall":31,"small":32,"medium":33},[],{"testimonial":191,"link":199,"type":175,"title":148,"description":176,"image":177},{"@type":120,"id":154,"model":117,"value":192},{"query":193,"folders":194,"createdDate":158,"id":154,"name":159,"modelId":127,"published":13,"data":195,"variations":196,"lastUpdated":167,"firstPublished":168,"testRatio":23,"createdBy":100,"lastUpdatedBy":24,"meta":197,"rev":171},[],[],{"video":161,"jobTitle":162,"author":163,"qoute":29,"quote":164,"image":165},{},{"kind":28,"lastPreviewUrl":29,"breakpoints":198,"hasAutosaves":34},{"small":32,"medium":33},{"text":173,"url":174},{},1776256937553,1776256937540,[],"xggsv10v1q9",{"createdDate":206,"id":207,"name":208,"modelId":186,"published":13,"stageModifiedSincePublish":6,"query":209,"data":210,"variations":220,"lastUpdated":221,"firstPublished":222,"testRatio":23,"createdBy":24,"lastUpdatedBy":24,"folders":223,"meta":224,"rev":204},1776256949234,"ce043785b71b4ece98eac811ecf4ba10","inductive-automation",[],{"link":211,"type":117,"testimonial":212,"testimonialLink":118},{},{"@type":120,"id":121,"model":117,"value":213},{"query":214,"folders":215,"createdDate":125,"id":121,"name":126,"modelId":127,"published":13,"data":216,"variations":217,"lastUpdated":133,"firstPublished":134,"testRatio":23,"createdBy":100,"lastUpdatedBy":100,"meta":218,"rev":137},[],[],{"author":129,"jobTitle":130,"quote":126,"image":131},{},{"kind":28,"lastPreviewUrl":29,"breakpoints":219,"hasAutosaves":34},{"small":32,"medium":33},{},1776256974140,1776256974130,[],{"breakpoints":225,"kind":28,"lastPreviewUrl":29,"hasAutosaves":6},{"xsmall":31,"small":32,"medium":33},{"id":227,"title":228,"authorsCollection":229,"content":237,"extension":877,"faqItemsCollection":878,"faqTitle":62,"featured":6,"hashTags":62,"meta":880,"metaTitle":881,"ogImage":62,"publishedDate":882,"relatedBlogPostsCollection":883,"slug":3396,"stem":3397,"subtitle":62,"summary":3398,"synopsis":3409,"sys":3410,"tagsCollection":3413,"__hash__":3419},"blog/blog/how-new-ai-agents-will-transform-credential-stuffing-attacks.json","How new AI agents will transform credential stuffing attacks",{"items":230},[231],{"fullName":232,"firstName":233,"jobTitle":234,"profilePicture":235},"Dan Green","Dan","Threat Research",{"url":236},"https://images.ctfassets.net/y1cdw1ablpvd/7jik1VhFgA3kgzXBXTm2Vw/fcd8c171da644903d0827eafcfbcaad0/Dan_Headshot_2025.png",{"json":238,"links":856},{"nodeType":239,"data":240,"content":241},"document",{},[242,253,275,308,315,319,327,334,343,350,357,364,371,379,386,415,444,451,459,466,491,498,501,509,516,524,544,577,584,596,603,611,618,625,632,639,642,650,657,680,689,692,700,716,723,726,734,741,758,765,795,798,805,825,844,850],{"nodeType":243,"data":244,"content":245},"heading-1",{},[246],{"nodeType":247,"value":248,"marks":249,"data":252},"text","Stolen credentials: The cyber criminal’s weapon of choice in 2024",[250],{"type":251},"bold",{},{"nodeType":254,"data":255,"content":256},"paragraph",{},[257,260,271],{"nodeType":247,"value":29,"marks":258,"data":259},[],{},{"nodeType":261,"data":262,"content":264},"hyperlink",{"uri":263},"https://www.verizon.com/business/resources/reports/dbir/",[265],{"nodeType":247,"value":266,"marks":267,"data":270},"Stolen credentials were the #1 attacker action in 2023/24",[268],{"type":269},"underline",{},{"nodeType":247,"value":272,"marks":273,"data":274},", and the breach vector for 80% of web app attacks. Not surprising when you consider the fact that billions of leaked credentials are in circulation online, and attackers can pick up the latest drop for as little as $10 on criminal forums. ",[],{},{"nodeType":254,"data":276,"content":277},{},[278,282,291,295,304],{"nodeType":247,"value":279,"marks":280,"data":281},"The criminal marketplace for stolen credentials is benefitting from ",[],{},{"nodeType":261,"data":283,"content":285},{"uri":284},"https://pushsecurity.com/resources/2024-identity-attacks",[286],{"nodeType":247,"value":287,"marks":288,"data":290},"the publicity of high profile breaches in 2024",[289],{"type":269},{},{"nodeType":247,"value":292,"marks":293,"data":294}," such as the attacks on ",[],{},{"nodeType":261,"data":296,"content":298},{"uri":297},"https://pushsecurity.com/blog/snowflake-retro/",[299],{"nodeType":247,"value":300,"marks":301,"data":303},"Snowflake",[302],{"type":269},{},{"nodeType":247,"value":305,"marks":306,"data":307}," customers using credentials found in data breach dumps and compromised credential feeds from infostealer and mass phishing campaigns, resulting in the compromise of 165 customer tenants and hundreds of millions of breached personal records.",[],{},{"nodeType":254,"data":309,"content":310},{},[311],{"nodeType":247,"value":312,"marks":313,"data":314},"But despite 2024 being an unprecedented year in terms of the impact of identity-based attacks, there’s still a lot of unfulfilled potential for attackers to realize. ",[],{},{"nodeType":316,"data":317,"content":318},"hr",{},[],{"nodeType":243,"data":320,"content":321},{},[322],{"nodeType":247,"value":323,"marks":324,"data":326},"Credential attack automation — what’s changed with the shift to SaaS? ",[325],{"type":251},{},{"nodeType":254,"data":328,"content":329},{},[330],{"nodeType":247,"value":331,"marks":332,"data":333},"Brute forcing and credential stuffing are nothing new, and have been a key component of the cyber attacker toolkit for decades. But it’s not quite as easy to automatically spray credentials across systems as it once was. ",[],{},{"nodeType":335,"data":336,"content":337},"heading-2",{},[338],{"nodeType":247,"value":339,"marks":340,"data":342},"No more one-size-fits-all",[341],{"type":251},{},{"nodeType":254,"data":344,"content":345},{},[346],{"nodeType":247,"value":347,"marks":348,"data":349},"Rather than a single centralized network with apps and data contained within an infrastructure perimeter, business IT is now formed of hundreds of web-based apps and platforms, creating thousands of identities per organization.  ",[],{},{"nodeType":254,"data":351,"content":352},{},[353],{"nodeType":247,"value":354,"marks":355,"data":356},"This means that identities too are now decentralized and distributed all over the internet, as opposed to being stored solely in identity systems like Active Directory, and implemented using common protocols and mechanisms. ",[],{},{"nodeType":254,"data":358,"content":359},{},[360],{"nodeType":247,"value":361,"marks":362,"data":363},"While HTTP(S) is standard, modern web apps are complex and highly customized, with a graphically-driven interface that is different every time. And to make matters worse, modern web apps are specifically designed to prevent malicious automation through bot protections like CAPTCHA. ",[],{},{"nodeType":254,"data":365,"content":366},{},[367],{"nodeType":247,"value":368,"marks":369,"data":370},"So rather than encountering standard protocols and being able to write a single set of tools to use across any organization/environment e.g. write a DNS scanner once, use a single port scanner like Nmap for the entire internet, write a single script per service (e.g. FTP, SSH, Telnet, etc.) for your password sprayer — custom tool development is instead required for every app that you want to target. ",[],{},{"nodeType":335,"data":372,"content":373},{},[374],{"nodeType":247,"value":375,"marks":376,"data":378},"Finding the needle in the haystack",[377],{"type":251},{},{"nodeType":254,"data":380,"content":381},{},[382],{"nodeType":247,"value":383,"marks":384,"data":385},"Not only are there more environments for attackers to include in the scope of their attack, but there are more credentials to work with. ",[],{},{"nodeType":254,"data":387,"content":388},{},[389,393,398,402,411],{"nodeType":247,"value":390,"marks":391,"data":392},"There are around ",[],{},{"nodeType":247,"value":394,"marks":395,"data":397},"15 billion compromised credentials",[396],{"type":251},{},{"nodeType":247,"value":399,"marks":400,"data":401}," available on the public internet, not including those found only in private channels/feeds. This list is growing all of the time — ",[],{},{"nodeType":261,"data":403,"content":405},{"uri":404},"https://www.bleepingcomputer.com/news/security/have-i-been-pwned-adds-284m-accounts-stolen-by-infostealer-malware/",[406],{"nodeType":247,"value":407,"marks":408,"data":410},"like 244M never before seen passwords and 493M unique website and email address pairs",[409],{"type":269},{},{"nodeType":247,"value":412,"marks":413,"data":414}," being added to Have I Been Pwned from infostealer logs just last month. ",[],{},{"nodeType":254,"data":416,"content":417},{},[418,422,431,435,440],{"nodeType":247,"value":419,"marks":420,"data":421},"This sounds scary, but it’s tricky for attackers to harness this data. The vast majority of these credentials are old and invalid. Our ",[],{},{"nodeType":261,"data":423,"content":425},{"uri":424},"https://pushsecurity.com/blog/verified-stolen-credential-detection/",[426],{"nodeType":247,"value":427,"marks":428,"data":430},"recent review of TI data",[429],{"type":269},{},{"nodeType":247,"value":432,"marks":433,"data":434}," found that fewer than 1% of stolen credentials included in threat intelligence feeds from a multi-vendor data set was actionable —",[],{},{"nodeType":247,"value":436,"marks":437,"data":439}," in other words, 99% of compromised credentials were false positives.",[438],{"type":251},{},{"nodeType":247,"value":441,"marks":442,"data":443}," ",[],{},{"nodeType":254,"data":445,"content":446},{},[447],{"nodeType":247,"value":448,"marks":449,"data":450},"But not all of them are useless — as the Snowflake attacks demonstrated, which successfully leveraged credentials dating back to 2020. So there are clearly treasures waiting to be discovered by attackers. ",[],{},{"nodeType":335,"data":452,"content":453},{},[454],{"nodeType":247,"value":455,"marks":456,"data":458},"Attackers are forced to prioritize",[457],{"type":251},{},{"nodeType":254,"data":460,"content":461},{},[462],{"nodeType":247,"value":463,"marks":464,"data":465},"The distributed nature of apps and identities, and the low reliability of compromised credential data, means attackers are forced to prioritize — despite a target-rich environment of hundreds of business apps, creating thousands of sprawled identities per organization, because: ",[],{},{"nodeType":467,"data":468,"content":469},"unordered-list",{},[470,481],{"nodeType":471,"data":472,"content":473},"list-item",{},[474],{"nodeType":254,"data":475,"content":476},{},[477],{"nodeType":247,"value":478,"marks":479,"data":480},"Writing and running custom python scripts for every single app (there are more than 40k SaaS apps on the internet) is not realistic. Even if you did the top 100 or 1000, that would be a significant task and require constant maintenance, while barely scratching the surface of the total opportunity. ",[],{},{"nodeType":471,"data":482,"content":483},{},[484],{"nodeType":254,"data":485,"content":486},{},[487],{"nodeType":247,"value":488,"marks":489,"data":490},"Even when fully scripted and using a botnet to distribute the attack and avoid IP blocking, controls like rate limiting, CAPTCHA, and account lockouts can obstruct mass credential stuffing against a single app. And a concentrated attack on a single site is going to generate significant levels of traffic if you want to get through 15 billion passwords in a reasonable timeframe, so it’s very likely to raise the alarm.",[],{},{"nodeType":254,"data":492,"content":493},{},[494],{"nodeType":247,"value":495,"marks":496,"data":497},"So attackers tend to target a smaller number of apps, and only look for a direct match in terms of the credentials attempted (e.g. the stolen credential must directly belong to an account on the target app). When they do go after something new, it tends to be concentrated on a specific app/platform (e.g. Snowflake) or looking for a narrower subset of credentials (e.g. credentials clearly associated with edge devices, for more traditional network environments). ",[],{},{"nodeType":316,"data":499,"content":500},{},[],{"nodeType":243,"data":502,"content":503},{},[504],{"nodeType":247,"value":505,"marks":506,"data":508},"A missed opportunity?",[507],{"type":251},{},{"nodeType":254,"data":510,"content":511},{},[512],{"nodeType":247,"value":513,"marks":514,"data":515},"As we’ve established, the situation regarding credential stuffing attacks is already pretty bad despite these limitations. But things could be significantly worse. ",[],{},{"nodeType":335,"data":517,"content":518},{},[519],{"nodeType":247,"value":520,"marks":521,"data":523},"Password reuse means a single compromised account could turn into many",[522],{"type":251},{},{"nodeType":254,"data":525,"content":526},{},[527,531,540],{"nodeType":247,"value":528,"marks":529,"data":530},"If attackers were able to increase the scale of their attacks to target a broader number of apps (rather than concentrating on a shortlist of high value apps) they could take advantage of all-too-common password reuse. According to our ",[],{},{"nodeType":261,"data":532,"content":534},{"uri":533},"https://pushsecurity.com/blog/how-many-vulnerable-identities-do-you-have/",[535],{"nodeType":247,"value":536,"marks":537,"data":539},"recent investigation of identity data",[538],{"type":269},{},{"nodeType":247,"value":541,"marks":542,"data":543},", on average:",[],{},{"nodeType":467,"data":545,"content":546},{},[547,557,567],{"nodeType":471,"data":548,"content":549},{},[550],{"nodeType":254,"data":551,"content":552},{},[553],{"nodeType":247,"value":554,"marks":555,"data":556},"1 in 3 employees reuse passwords",[],{},{"nodeType":471,"data":558,"content":559},{},[560],{"nodeType":254,"data":561,"content":562},{},[563],{"nodeType":247,"value":564,"marks":565,"data":566},"9% of identities have a reused password AND no MFA",[],{},{"nodeType":471,"data":568,"content":569},{},[570],{"nodeType":254,"data":571,"content":572},{},[573],{"nodeType":247,"value":574,"marks":575,"data":576},"10% of IdP accounts (used for SSO) have a non-unique password",[],{},{"nodeType":254,"data":578,"content":579},{},[580],{"nodeType":247,"value":581,"marks":582,"data":583},"What does this mean? If a stolen credential is valid, there’s a good chance that it can be used to access more than one account, on more than one app (at least). ",[],{},{"nodeType":254,"data":585,"content":586},{},[587,592],{"nodeType":247,"value":588,"marks":589,"data":591},"Picture the scenario:",[590],{"type":251},{},{"nodeType":247,"value":593,"marks":594,"data":595}," A recent compromised credential leak from infostealer infections or credential phishing campaigns shows that a particular username and password combination is valid on a specific app — let’s say Microsoft 365. Now, this account is pretty locked down — not only does it have MFA, but there are conditional access policies in place restricting the IP/location it can be accessed from. ",[],{},{"nodeType":254,"data":597,"content":598},{},[599],{"nodeType":247,"value":600,"marks":601,"data":602},"Usually, this is where the attack would end, and you’d turn your attention to something else. But what if you were able to spray these credentials across every other business app that the user has an account on?",[],{},{"nodeType":335,"data":604,"content":605},{},[606],{"nodeType":247,"value":607,"marks":608,"data":610},"Scaling credential attacks with Computer-Using Agents",[609],{"type":251},{},{"nodeType":254,"data":612,"content":613},{},[614],{"nodeType":247,"value":615,"marks":616,"data":617},"Until now, the impact of AI on identity attacks has been limited to the use of LLMs for the creation of phishing emails, in AI-assisted malware development, and for social media bots — no doubt significant, but not exactly transformative, and requiring constant human oversight and input. ",[],{},{"nodeType":254,"data":619,"content":620},{},[621],{"nodeType":247,"value":622,"marks":623,"data":624},"But with the launch of OpenAI Operator, a new kind of “Computer-Using Agent”, this could be about to change. ",[],{},{"nodeType":254,"data":626,"content":627},{},[628],{"nodeType":247,"value":629,"marks":630,"data":631},"Operator is trained on a specialist dataset and implemented in its own sandboxed browser, meaning it is able to perform common web tasks like a human — seeing and interacting with pages as a human would. ",[],{},{"nodeType":254,"data":633,"content":634},{},[635],{"nodeType":247,"value":636,"marks":637,"data":638},"Unlike other automated solutions, Operator requires no custom implementation or coding to be able to interact with new sites, making it a much more scalable option for attackers looking to target a broad sweep of sites/apps. ",[],{},{"nodeType":316,"data":640,"content":641},{},[],{"nodeType":243,"data":643,"content":644},{},[645],{"nodeType":247,"value":646,"marks":647,"data":649},"Demo: Using Operator to conduct credential stuffing attacks at scale",[648],{"type":251},{},{"nodeType":254,"data":651,"content":652},{},[653],{"nodeType":247,"value":654,"marks":655,"data":656},"We put the malicious use-cases of Operator to the test, using it to:",[],{},{"nodeType":467,"data":658,"content":659},{},[660,670],{"nodeType":471,"data":661,"content":662},{},[663],{"nodeType":254,"data":664,"content":665},{},[666],{"nodeType":247,"value":667,"marks":668,"data":669},"Identify which companies have an existing tenant on a list of apps",[],{},{"nodeType":471,"data":671,"content":672},{},[673],{"nodeType":254,"data":674,"content":675},{},[676],{"nodeType":247,"value":677,"marks":678,"data":679},"Attempt to login to various app tenants with a provided username and password",[],{},{"nodeType":681,"data":682,"content":688},"embedded-entry-block",{"target":683},{"sys":684},{"id":685,"type":686,"linkType":687},"moAA7jwcBmZMaqpmPva4w","Link","Entry",[],{"nodeType":316,"data":690,"content":691},{},[],{"nodeType":243,"data":693,"content":694},{},[695],{"nodeType":247,"value":696,"marks":697,"data":699},"Impact summary",[698],{"type":251},{},{"nodeType":254,"data":701,"content":702},{},[703,707,712],{"nodeType":247,"value":704,"marks":705,"data":706},"The results were pretty eye opening. Operator clearly demonstrated the ability to target a list of apps with compromised credentials and perform in-app actions. ",[],{},{"nodeType":247,"value":708,"marks":709,"data":711},"Now think about this x10, x100, x10,000 …",[710],{"type":251},{},{"nodeType":247,"value":713,"marks":714,"data":715}," These are not complex tasks. But the value of CUAs Operator is not in tackling complexity, but scale. Imagine a world where you can orchestrate Operator windows via API and get it to execute these actions simultaneously (functionality that exists already for ChatGPT). ",[],{},{"nodeType":254,"data":717,"content":718},{},[719],{"nodeType":247,"value":720,"marks":721,"data":722},"But this is bigger than Operator — it’s about the direction of the technology. OpenAI may implement restrictions — better in-app guardrails, rate limits on the number of concurrent tasks and total usage, etc. But you can guarantee it won’t be the only CUA — it’s only a matter of time before similar products emerge (maybe even inherently malicious ones) making use of the same technology. ",[],{},{"nodeType":316,"data":724,"content":725},{},[],{"nodeType":243,"data":727,"content":728},{},[729],{"nodeType":247,"value":730,"marks":731,"data":733},"Final thoughts ",[732],{"type":251},{},{"nodeType":254,"data":735,"content":736},{},[737],{"nodeType":247,"value":738,"marks":739,"data":740},"It’s still early days for CUA tech, but there’s a clear indication that an already severe security challenge could be made worse with this particular form of AI-driven automation. While the ability to target a broad set of apps has been previously beyond the scope of traditional automation, it’s about to become much more accessible to even low-skilled attackers (think: next gen script kiddies?). ",[],{},{"nodeType":254,"data":742,"content":743},{},[744,748,754],{"nodeType":247,"value":745,"marks":746,"data":747},"Another way to think about it is that it effectively gives a human attacker a fleet of low-level interns who don’t ",[],{},{"nodeType":247,"value":749,"marks":750,"data":753},"quite",[751],{"type":752},"italic",{},{"nodeType":247,"value":755,"marks":756,"data":757}," know what they’re doing, but can be instructed to perform specific, itemised tasks at scale with only the occasional check in — while you work on other, more complex tasks. So, a bit like a red team manager of AI bots. ",[],{},{"nodeType":254,"data":759,"content":760},{},[761],{"nodeType":247,"value":762,"marks":763,"data":764},"Operator means that attackers can leverage compromised credentials at-scale, take advantage of the vast numbers of vulnerable and misconfigured identities, and convert them into systemic breaches much more easily. In a way, it could make credential stuffing a bit more like it was before the shift to cloud apps — where you could spray thousands of credentials across your targets without needing custom development every time. ",[],{},{"nodeType":254,"data":766,"content":767},{},[768,772,779,783,791],{"nodeType":247,"value":769,"marks":770,"data":771},"Thankfully, no new anti-AI capabilities are required — but it’s more important than ever that organizations look to defend their identity attack surface and find and fix identity vulnerabilities before attackers can take advantage of them. To prevent stolen credentials being exploited, defenders need to be able ",[],{},{"nodeType":261,"data":773,"content":774},{"uri":424},[775],{"nodeType":247,"value":776,"marks":777,"data":778},"to quickly identify where compromised credentials are being actively used",[],{},{"nodeType":247,"value":780,"marks":781,"data":782},", and ",[],{},{"nodeType":261,"data":784,"content":786},{"uri":785},"https://pushsecurity.com/blog/enforce-mfa-on-third-party-apps/",[787],{"nodeType":247,"value":788,"marks":789,"data":790},"deploy mitigating controls like MFA",[],{},{"nodeType":247,"value":792,"marks":793,"data":794}," to prevent exploitation with a vulnerable single factor. ",[],{},{"nodeType":316,"data":796,"content":797},{},[],{"nodeType":243,"data":799,"content":800},{},[801],{"nodeType":247,"value":802,"marks":803,"data":804},"How Push can help",[],{},{"nodeType":254,"data":806,"content":807},{},[808,812,821],{"nodeType":247,"value":809,"marks":810,"data":811},"If you want to learn more about identity attacks and how to stop them, ",[],{},{"nodeType":261,"data":813,"content":815},{"uri":814},"https://pushsecurity.com/demo",[816],{"nodeType":247,"value":817,"marks":818,"data":820},"book a demo",[819],{"type":269},{},{"nodeType":247,"value":822,"marks":823,"data":824},".",[],{},{"nodeType":254,"data":826,"content":827},{},[828,832,841],{"nodeType":247,"value":829,"marks":830,"data":831},"And if you want to see more malicious use cases of Operator in action, ",[],{},{"nodeType":261,"data":833,"content":835},{"uri":834},"https://pushsecurity.com/resources/automating-identity-attacks",[836],{"nodeType":247,"value":837,"marks":838,"data":840},"check out this on-demand webinar. ",[839],{"type":269},{},{"nodeType":247,"value":29,"marks":842,"data":843},[],{},{"nodeType":681,"data":845,"content":849},{"target":846},{"sys":847},{"id":848,"type":686,"linkType":687},"6169dNBRXvahtV8VRlxLCJ",[],{"nodeType":254,"data":851,"content":852},{},[853],{"nodeType":247,"value":29,"marks":854,"data":855},[],{},{"entries":857},{"hyperlink":858,"inline":859,"block":860},[],[],[861,870],{"sys":862,"__typename":863,"title":864,"youTubeUrl":865,"imagePlaceholder":866},{"id":685},"ExternalVideo","Using Operator to conduct credential stuffing attacks at scale","https://www.youtube.com/watch?v=a_YJafxPjMo",{"url":867,"width":868,"height":869},"https://images.ctfassets.net/y1cdw1ablpvd/47KUHgOYv3FHVkGUbwA3Ly/ab5d478e6b5b1fff84248a05b6d3ccaf/Operator_clip_thumbnail__3_.jpg",1920,1080,{"sys":871,"__typename":872,"type":873,"ctaText":874,"buttonLabel":875,"buttonColour":876,"buttonUrl":834},{"id":848},"CtaWidget","Custom","See our latest webinar on demand where we explore Operator's malicious use-cases throughout the kill chain","Watch Now","sea blue","json",{"items":879},[],{},"Using OpenAI Operator to automate credential stuffing","2025-03-11T00:00:00.000Z",{"items":884},[885,1531,2735],{"__typename":886,"sys":887,"content":889,"title":1509,"synopsis":1510,"hashTags":62,"publishedDate":1511,"slug":1512,"tagsCollection":1513,"authorsCollection":1523},"BlogPosts",{"id":888},"SiALa9w13C6q3OzeTeUum",{"json":890},{"nodeType":239,"data":891,"content":892},{},[893,900,916,923,929,936,943,946,953,960,967,991,998,1014,1017,1024,1031,1054,1061,1068,1087,1093,1100,1119,1125,1132,1178,1186,1193,1205,1217,1224,1257,1264,1267,1274,1294,1301,1308,1315,1318,1325,1333,1340,1347,1410,1417,1424,1431,1464,1470,1477,1484,1490],{"nodeType":254,"data":894,"content":895},{},[896],{"nodeType":247,"value":897,"marks":898,"data":899},"Computer-Using Agents (CUAs) are a new type of AI agent that drives your browser/OS for you. With the research preview release of OpenAI Operator last week, it’s likely that we’ll be seeing a lot more of this technology in the future as OpenAI iterates and competitors launch their own versions. ",[],{},{"nodeType":254,"data":901,"content":902},{},[903,907,912],{"nodeType":247,"value":904,"marks":905,"data":906},"These models run on the same UI as the user sees, rather than using code or API based add-ons or tools (e.g. with access via API keys). In Operator’s case, the agent runs in its own browser, where it can navigate to and interact with webpages by typing, clicking, and scrolling. It effectively sees and interacts with pages as a human would, ",[],{},{"nodeType":247,"value":908,"marks":909,"data":911},"using human (not machine) identities",[910],{"type":251},{},{"nodeType":247,"value":913,"marks":914,"data":915}," — taking actions on the web without requiring custom API integrations. ",[],{},{"nodeType":254,"data":917,"content":918},{},[919],{"nodeType":247,"value":920,"marks":921,"data":922},"This means that a user describes a task, and Operator performs it autonomously on their behalf. The examples provided by OpenAI are things like booking a dinner reservation or shopping for groceries — but naturally the potential use cases are much, much broader, especially in a work context.",[],{},{"nodeType":681,"data":924,"content":928},{"target":925},{"sys":926},{"id":927,"type":686,"linkType":687},"5mWWi5mfqEcSQX12gOtyQm",[],{"nodeType":254,"data":930,"content":931},{},[932],{"nodeType":247,"value":933,"marks":934,"data":935},"Obviously the broad impact of this technology is almost impossible to predict this early in the game. But since we’re focussed on identity security at Push, we can at least describe some of the very predictable impacts in this area.",[],{},{"nodeType":254,"data":937,"content":938},{},[939],{"nodeType":247,"value":940,"marks":941,"data":942},"CUAs like Operator are essentially very flexible no-code automation platforms. This means that these tools (or future iterations of them) will enable low-cost, low-effort automation of common web tasks — the very tasks that app developers and vendors have worked hard to prevent from being automated — including those frequently performed by attackers.",[],{},{"nodeType":316,"data":944,"content":945},{},[],{"nodeType":243,"data":947,"content":948},{},[949],{"nodeType":247,"value":950,"marks":951,"data":952},"Why do CUAs stand to benefit attackers more than previous AI tools? ",[],{},{"nodeType":254,"data":954,"content":955},{},[956],{"nodeType":247,"value":957,"marks":958,"data":959},"Organizations have been concerned about the security and privacy implications of GenAI tools and platforms for a while now — mainly concerning the risk of inputting sensitive data into LLMs, and prompt injection attacks in which models can be tricked into disclosing internal data. ",[],{},{"nodeType":254,"data":961,"content":962},{},[963],{"nodeType":247,"value":964,"marks":965,"data":966},"But so far, the primary impact of GenAI on attacker capabilities specifically has been mainly limited to the use of LLMs for the creation of phishing emails and in AI-assisted malware development — no doubt significant, but not exactly transformative. And although the concept of an AI agent is nothing new, they haven’t been particularly common outside of research circles. ",[],{},{"nodeType":254,"data":968,"content":969},{},[970,974,979,983,988],{"nodeType":247,"value":971,"marks":972,"data":973},"CUAs, on the other hand, use LLMs trained using datasets which make them far more able to understand and interact with web pages. Coupled with what is essentially a production-grade integration between browser and LLM, and you have an agent that is able to understand and interact with websites to achieve an outcome, with minimal human input and oversight (as opposed to simply scraping the data) ",[],{},{"nodeType":247,"value":975,"marks":976,"data":978},"with much the same behaviors and capabilities",[977],{"type":251},{},{"nodeType":247,"value":980,"marks":981,"data":982}," ",[],{},{"nodeType":247,"value":984,"marks":985,"data":987},"as a human operator.",[986],{"type":251},{},{"nodeType":247,"value":441,"marks":989,"data":990},[],{},{"nodeType":254,"data":992,"content":993},{},[994],{"nodeType":247,"value":995,"marks":996,"data":997},"By performing actions autonomously on the user’s behalf, it has a lot in common with a low/no-code automation platform like Zapier or Make.com — except it doesn’t perform actions via API, but by performing actions in the browser as a user would. Unlike no/low-code automations, it doesn’t need a strict or rigid step-by-step description of tasks that should be automated and can dynamically generate steps like a human does. ",[],{},{"nodeType":254,"data":999,"content":1000},{},[1001,1005,1010],{"nodeType":247,"value":1002,"marks":1003,"data":1004},"None of this can’t be done using other automation tools, but it’s the difference between writing code to automate a task by hand and asking a human assistant to do something for you — ",[],{},{"nodeType":247,"value":1006,"marks":1007,"data":1009},"the effort required is reduced by orders of magnitude.",[1008],{"type":251},{},{"nodeType":247,"value":1011,"marks":1012,"data":1013}," This makes it both more flexible and accessible to a much wider range of users. ",[],{},{"nodeType":316,"data":1015,"content":1016},{},[],{"nodeType":243,"data":1018,"content":1019},{},[1020],{"nodeType":247,"value":1021,"marks":1022,"data":1023},"How can CUAs be abused by attackers?",[],{},{"nodeType":254,"data":1025,"content":1026},{},[1027],{"nodeType":247,"value":1028,"marks":1029,"data":1030},"There are two main groups of attack to be aware of:",[],{},{"nodeType":467,"data":1032,"content":1033},{},[1034,1044],{"nodeType":471,"data":1035,"content":1036},{},[1037],{"nodeType":254,"data":1038,"content":1039},{},[1040],{"nodeType":247,"value":1041,"marks":1042,"data":1043},"Attacks enabled by the technology (CUA)",[],{},{"nodeType":471,"data":1045,"content":1046},{},[1047],{"nodeType":254,"data":1048,"content":1049},{},[1050],{"nodeType":247,"value":1051,"marks":1052,"data":1053},"Attacks against specific CUA tools/implementations (e.g. Operator)",[],{},{"nodeType":254,"data":1055,"content":1056},{},[1057],{"nodeType":247,"value":1058,"marks":1059,"data":1060},"Because the answer to the latter question is subjective depending on the CUA being targeted (and Operator is still in its “research preview” release) we’ll focus on how attackers can potentially use CUAs for malicious purposes in general. ",[],{},{"nodeType":335,"data":1062,"content":1063},{},[1064],{"nodeType":247,"value":1065,"marks":1066,"data":1067},"How attackers can use their own CUAs to conduct AI-powered cyber attacks",[],{},{"nodeType":254,"data":1069,"content":1070},{},[1071,1075,1083],{"nodeType":247,"value":1072,"marks":1073,"data":1074},"The most obvious use-case for an attacker-controlled CUA is targeting internet-based app accounts. Most organizations are now using hundreds of apps, with thousands of sprawling identities (including both inside enterprise SSO connected accounts and local username & password logins) — ",[],{},{"nodeType":261,"data":1076,"content":1077},{"uri":533},[1078],{"nodeType":247,"value":1079,"marks":1080,"data":1082},"many of which are highly vulnerable to even low-sophistication attack techniques",[1081],{"type":269},{},{"nodeType":247,"value":1084,"marks":1085,"data":1086},". ",[],{},{"nodeType":681,"data":1088,"content":1092},{"target":1089},{"sys":1090},{"id":1091,"type":686,"linkType":687},"7itjimRwqpkrCF7YRI8FTq",[],{"nodeType":254,"data":1094,"content":1095},{},[1096],{"nodeType":247,"value":1097,"marks":1098,"data":1099},"Previously, identity attacks against modern SaaS environments and the sprawl of apps and accounts required a lot of manual work to scale. Because web identities are implemented in mostly bespoke ways across thousands of sites (and they are constantly changing) attacks on them are challenging to automate. Further, the act of logging in using automated methods has been impacted by widespread bot protection — specifically to prevent malicious automation. ",[],{},{"nodeType":254,"data":1101,"content":1102},{},[1103,1107,1115],{"nodeType":247,"value":1104,"marks":1105,"data":1106},"So, attackers end up sending phishing links through email, and targeting only a few high value apps for cred stuffing — despite the availability of credentials online (which, ",[],{},{"nodeType":261,"data":1108,"content":1109},{"uri":297},[1110],{"nodeType":247,"value":1111,"marks":1112,"data":1114},"as the Snowflake attacks demonstrate",[1113],{"type":269},{},{"nodeType":247,"value":1116,"marks":1117,"data":1118},", can be an untapped treasure trove for attackers).",[],{},{"nodeType":681,"data":1120,"content":1124},{"target":1121},{"sys":1122},{"id":1123,"type":686,"linkType":687},"24HV5O6LJ12ZVECTSel2WL",[],{"nodeType":254,"data":1126,"content":1127},{},[1128],{"nodeType":247,"value":1129,"marks":1130,"data":1131},"We know that about 1 in 3 users re-use passwords, so there is a great chance a lot of those exact same credentials were actually valid for many other apps. It’s very tough to manually test each credential by logging into even a few dozen apps (or building a web automation to do so). But this is significantly easier if you can ask a CUA to: ",[],{},{"nodeType":467,"data":1133,"content":1134},{},[1135,1145,1155],{"nodeType":471,"data":1136,"content":1137},{},[1138],{"nodeType":254,"data":1139,"content":1140},{},[1141],{"nodeType":247,"value":1142,"marks":1143,"data":1144},"“Find a list of the top 1000 SaaS apps”. ",[],{},{"nodeType":471,"data":1146,"content":1147},{},[1148],{"nodeType":254,"data":1149,"content":1150},{},[1151],{"nodeType":247,"value":1152,"marks":1153,"data":1154},"“Try to login to the app using this username and password. Let me know which apps you successfully logged into”. ",[],{},{"nodeType":471,"data":1156,"content":1157},{},[1158],{"nodeType":254,"data":1159,"content":1160},{},[1161,1165,1174],{"nodeType":247,"value":1162,"marks":1163,"data":1164},"“Use ",[],{},{"nodeType":261,"data":1166,"content":1168},{"uri":1167},"https://github.com/pushsecurity/saas-attacks/blob/main/techniques/takeout_services/description.md",[1169],{"nodeType":247,"value":1170,"marks":1171,"data":1173},"takeout services",[1172],{"type":269},{},{"nodeType":247,"value":1175,"marks":1176,"data":1177}," to download data from each app and send it to this location, grouping by company name” (or even just ask the model to cut and paste or download the data from the account).",[],{},{"nodeType":254,"data":1179,"content":1180},{},[1181],{"nodeType":247,"value":1182,"marks":1183,"data":1185},"This is how you really scale these attacks.",[1184],{"type":251},{},{"nodeType":254,"data":1187,"content":1188},{},[1189],{"nodeType":247,"value":1190,"marks":1191,"data":1192},"CUA agents also change how and where phishing can take place. Where phishing takes place outside of email, it’s much less likely to be intercepted by enterprise anti-phishing controls. You could:",[],{},{"nodeType":254,"data":1194,"content":1195},{},[1196,1201],{"nodeType":247,"value":1197,"marks":1198,"data":1200},"1.",[1199],{"type":251},{},{"nodeType":247,"value":1202,"marks":1203,"data":1204}," Task an agent to create Reddit, Discord, and Slack accounts, login, and find the 100 (or 10000?) biggest subreddits/communities/channels. Now have it join those, and write posts that seem relevant to ongoing threads, or write targeted DMs and include links to a phishing page. If the account gets banned, no problem, automatically start over. Not enough karma? Instruct the agent to build karma.",[],{},{"nodeType":254,"data":1206,"content":1207},{},[1208,1213],{"nodeType":247,"value":1209,"marks":1210,"data":1212},"2.",[1211],{"type":251},{},{"nodeType":247,"value":1214,"marks":1215,"data":1216}," Or consider a more targeted scenario: connect to a specific target (or group of targets) via LinkedIn, read all your target’s posts and comments, and using that context start a conversation with them, using a topic you know that will interest them to create a phishing lure, and direct them to your phishing site. ",[],{},{"nodeType":335,"data":1218,"content":1219},{},[1220],{"nodeType":247,"value":1221,"marks":1222,"data":1223},"Operator caveats",[],{},{"nodeType":254,"data":1225,"content":1226},{},[1227,1231,1240,1244,1253],{"nodeType":247,"value":1228,"marks":1229,"data":1230},"Now, it’s worth pointing out that Operator has controls that are designed to prevent this sort of abuse. ",[],{},{"nodeType":261,"data":1232,"content":1234},{"uri":1233},"https://openai.com/index/introducing-operator/",[1235],{"nodeType":247,"value":1236,"marks":1237,"data":1239},"For example",[1238],{"type":269},{},{"nodeType":247,"value":1241,"marks":1242,"data":1243},", Operator is trained to proactively ask the user to take over for tasks that require login, payment details, or when solving CAPTCHAs. The ",[],{},{"nodeType":261,"data":1245,"content":1247},{"uri":1246},"https://openai.com/index/operator-system-card/",[1248],{"nodeType":247,"value":1249,"marks":1250,"data":1252},"Operator System Card",[1251],{"type":269},{},{"nodeType":247,"value":1254,"marks":1255,"data":1256}," also cites proactive refusals of high-risk tasks, confirmation prompts before critical actions, and active monitoring systems to detect and mitigate potential threats.",[],{},{"nodeType":254,"data":1258,"content":1259},{},[1260],{"nodeType":247,"value":1261,"marks":1262,"data":1263},"It’s unclear at this point how resistant Operator will be to attack or abuse, but really, as we said earlier, this is not about Operator — once CUA tech becomes more widely available (if recent trends are anything to go by) there’s no doubt that models will emerge with fewer (or no) safety controls. ",[],{},{"nodeType":316,"data":1265,"content":1266},{},[],{"nodeType":243,"data":1268,"content":1269},{},[1270],{"nodeType":247,"value":1271,"marks":1272,"data":1273},"Why CUA-based automation is a problem for security teams",[],{},{"nodeType":254,"data":1275,"content":1276},{},[1277,1281,1290],{"nodeType":247,"value":1278,"marks":1279,"data":1280},"Attackers have been using automation tools forever, and in response, developers have been building protections against them (e.g. Cloudflare Turnstile and CAPTCHAs). Using LLMs to super power them isn’t even new, nor is using automation apps for malicious purposes (see our SaaS attack matrix entry for ",[],{},{"nodeType":261,"data":1282,"content":1284},{"uri":1283},"https://github.com/pushsecurity/saas-attacks/blob/main/techniques/shadow_workflows/description.md",[1285],{"nodeType":247,"value":1286,"marks":1287,"data":1289},"shadow workflows",[1288],{"type":269},{},{"nodeType":247,"value":1291,"marks":1292,"data":1293},") — so what’s the difference?",[],{},{"nodeType":254,"data":1295,"content":1296},{},[1297],{"nodeType":247,"value":1298,"marks":1299,"data":1300},"Previously, attackers needed to tie together automated browsers, get bot protection bypasses working, write code to extract screenshots from these browsers, pump those screenshots into a traditional LLM, generate response actions, and write code to execute those actions using browser automation. It was a lot of manual work — and needed constant maintenance — and wasn’t very effective because the general LLMs weren’t good at interpreting what they were seeing.",[],{},{"nodeType":254,"data":1302,"content":1303},{},[1304],{"nodeType":247,"value":1305,"marks":1306,"data":1307},"So, this isn’t so much a change in capability but a signal that there is going to be a massive increase in performance compared to other AI agents. Bundle the new model’s ability to understand with the ability to interact with webpages and you have something that might soon create real world impact at scale. ",[],{},{"nodeType":254,"data":1309,"content":1310},{},[1311],{"nodeType":247,"value":1312,"marks":1313,"data":1314},"Perhaps the only real obstacles are safety controls and cost. But as we’ve seen after previous GenAI launches, most recently with DeepSeek — competitors have been fast following with models that out-perform the original. Some of these models will be open and contain far fewer safety protections. An open CUA model in the future might be the trigger that enables attackers to leverage these capabilities at scale. ",[],{},{"nodeType":316,"data":1316,"content":1317},{},[],{"nodeType":243,"data":1319,"content":1320},{},[1321],{"nodeType":247,"value":1322,"marks":1323,"data":1324},"So what?",[],{},{"nodeType":254,"data":1326,"content":1327},{},[1328],{"nodeType":247,"value":1329,"marks":1330,"data":1332},"The TL;DR is that the adoption of CUAs has the potential to significantly lower the cost to attackers of running identity attacks such as phishing and credential stuffing, while increasing their reach.",[1331],{"type":251},{},{"nodeType":254,"data":1334,"content":1335},{},[1336],{"nodeType":247,"value":1337,"marks":1338,"data":1339},"We can expect improved account takeover attacks in the future as this technology becomes more widespread, with phishing attacks being increasingly delivered outside of traditional (well-protected) mediums like email, and credential stuffing being weaponized on an even more widespread scale, across a broader range of apps. These capabilities will also become more accessible, with even less advanced attackers able to harness them.",[],{},{"nodeType":254,"data":1341,"content":1342},{},[1343],{"nodeType":247,"value":1344,"marks":1345,"data":1346},"Right now, Operator runs in a sandboxed browser environment. But going forward, more value will require an increased ability to perform authenticated access as the user — so one could imagine a world where new features are built to expose passwords into this sandbox — or that these agents will be enabled outside these sandboxes and operate in your browser (primarily) or directly on your OS using agents. We’ve already seen these agents implemented as browser extensions. This makes sense as extensions can see the tab, and interact with the page — and some early extension-based agents have existed for a while:",[],{},{"nodeType":467,"data":1348,"content":1349},{},[1350,1370,1390],{"nodeType":471,"data":1351,"content":1352},{},[1353],{"nodeType":254,"data":1354,"content":1355},{},[1356,1359,1367],{"nodeType":247,"value":29,"marks":1357,"data":1358},[],{},{"nodeType":261,"data":1360,"content":1362},{"uri":1361},"https://github.com/richardyc/Chrome-GPT",[1363],{"nodeType":247,"value":1361,"marks":1364,"data":1366},[1365],{"type":269},{},{"nodeType":247,"value":441,"marks":1368,"data":1369},[],{},{"nodeType":471,"data":1371,"content":1372},{},[1373],{"nodeType":254,"data":1374,"content":1375},{},[1376,1379,1387],{"nodeType":247,"value":29,"marks":1377,"data":1378},[],{},{"nodeType":261,"data":1380,"content":1382},{"uri":1381},"https://github.com/handrew/browserpilot",[1383],{"nodeType":247,"value":1381,"marks":1384,"data":1386},[1385],{"type":269},{},{"nodeType":247,"value":29,"marks":1388,"data":1389},[],{},{"nodeType":471,"data":1391,"content":1392},{},[1393],{"nodeType":254,"data":1394,"content":1395},{},[1396,1399,1407],{"nodeType":247,"value":29,"marks":1397,"data":1398},[],{},{"nodeType":261,"data":1400,"content":1402},{"uri":1401},"https://github.com/TaxyAI/browser-extension",[1403],{"nodeType":247,"value":1401,"marks":1404,"data":1406},[1405],{"type":269},{},{"nodeType":247,"value":822,"marks":1408,"data":1409},[],{},{"nodeType":254,"data":1411,"content":1412},{},[1413],{"nodeType":247,"value":1414,"marks":1415,"data":1416},"If we have agents operating on user endpoints, not in sandboxes, that means they will have access to all identities that are already authenticated, or that can be automatically authenticated (password manager autofills etc.). There’s nothing fundamentally stopping you from prompt-injecting a victim's CUA and tricking it into creating a malicious integration, or sending you an API key.",[],{},{"nodeType":335,"data":1418,"content":1419},{},[1420],{"nodeType":247,"value":1421,"marks":1422,"data":1423},"So to summarize...",[],{},{"nodeType":254,"data":1425,"content":1426},{},[1427],{"nodeType":247,"value":1428,"marks":1429,"data":1430},"Organizations should anticipate an increase in identity attacks targeting web-based apps and services using techniques that can be amplified by CUAs such as phishing and credential stuffing. We recommend that organizations:",[],{},{"nodeType":467,"data":1432,"content":1433},{},[1434,1444,1454],{"nodeType":471,"data":1435,"content":1436},{},[1437],{"nodeType":254,"data":1438,"content":1439},{},[1440],{"nodeType":247,"value":1441,"marks":1442,"data":1443},"Anticipate an increase in phishing attacks delivered outside of email, and evaluate your detection capabilities for mediums such as IM platforms and social media sites.",[],{},{"nodeType":471,"data":1445,"content":1446},{},[1447],{"nodeType":254,"data":1448,"content":1449},{},[1450],{"nodeType":247,"value":1451,"marks":1452,"data":1453},"Find and harden identities that could be vulnerable to attacks using techniques that can be automated (e.g. mass credential stuffing) such as those missing phishing resistant MFA (or MFA altogether).",[],{},{"nodeType":471,"data":1455,"content":1456},{},[1457],{"nodeType":254,"data":1458,"content":1459},{},[1460],{"nodeType":247,"value":1461,"marks":1462,"data":1463},"Ensure that all identities are suitably protected — even those outside the scope of traditional identity stores (such as Active Directory and modern equivalents e.g. Entra, Okta) used to access the much broader set of web-based services. ",[],{},{"nodeType":335,"data":1465,"content":1466},{},[1467],{"nodeType":247,"value":802,"marks":1468,"data":1469},[],{},{"nodeType":254,"data":1471,"content":1472},{},[1473],{"nodeType":247,"value":1474,"marks":1475,"data":1476},"AI-powered or not, identity attacks are what Push is designed to combat. Our features and controls designed to stop account takeover via phishing, credential stuffing, and session hijacking remain effective in this new world — in fact, as attackers are granted the ability to conduct these attacks with greater speed and scale, they become more valuable than ever. ",[],{},{"nodeType":254,"data":1478,"content":1479},{},[1480],{"nodeType":247,"value":1481,"marks":1482,"data":1483},"If you're interested in learning more, check out our on-demand webinar where we demonstrate the use of CUAs for automating identity attacks, particularly in the context of SaaS account takeover. ",[],{},{"nodeType":681,"data":1485,"content":1489},{"target":1486},{"sys":1487},{"id":1488,"type":686,"linkType":687},"UCmd5kqVZ03ce5Cs9M0r5",[],{"nodeType":254,"data":1491,"content":1492},{},[1493,1497,1505],{"nodeType":247,"value":1494,"marks":1495,"data":1496},"If you’d like to learn more about Push, ",[],{},{"nodeType":261,"data":1498,"content":1500},{"uri":1499},"https://pushsecurity.com/demo/",[1501],{"nodeType":247,"value":1502,"marks":1503,"data":1504},"set up a demo with our team",[],{},{"nodeType":247,"value":1506,"marks":1507,"data":1508}," or sign up yourself to have a look at the platform.",[],{},"Considering the security implications of Computer-Using Agents (like OpenAI Operator)","CUAs are a new type of AI agent that drives your browser/OS for you, enabling effortless automation of web tasks — including those performed by attackers.","2025-01-28T00:00:00.000Z","considering-the-impact-of-computer-using-agents",{"items":1514},[1515,1519],{"sys":1516,"name":1518},{"id":1517},"6A5RXS31ZQx3PwryGb1IMy","Browser-based attacks",{"sys":1520,"name":1522},{"id":1521},"3pjES4THCIfSAwhGdNwBcy","Browser security",{"items":1524},[1525],{"fullName":1526,"firstName":1527,"jobTitle":1528,"profilePicture":1529},"Jacques Louw","Jacques","Co-founder / CRO",{"url":1530},"https://images.ctfassets.net/y1cdw1ablpvd/39m8bektV23lnCRcEq0G8h/2a08f6276a50744f1a4b499b273f6bb2/Push_Founders_at_Cahoots_October_28_2022_by_Doug_Coombe-21.jpg",{"__typename":886,"sys":1532,"content":1534,"title":2715,"synopsis":2716,"hashTags":62,"publishedDate":2717,"slug":2718,"tagsCollection":2719,"authorsCollection":2727},{"id":1533},"5KqYY7p174lSpuinfTfEZU",{"json":1535},{"data":1536,"content":1537,"nodeType":239},{},[1538,1545,1600,1607,1610,1617,1624,1657,1669,1672,1679,1690,1697,1717,1723,1743,1750,1761,1782,1802,1809,1828,1834,1854,1857,1864,1871,1878,1898,1918,1925,1945,1952,1959,1984,1991,2011,2032,2038,2045,2065,2068,2075,2094,2113,2120,2173,2180,2183,2190,2197,2216,2236,2243,2250,2257,2278,2285,2292,2298,2304,2307,2314,2321,2328,2348,2359,2378,2385,2392,2402,2409,2416,2436,2442,2445,2452,2459,2466,2539,2546,2553,2560,2568,2589,2596,2602,2613,2634,2641,2649,2669,2676,2683,2689,2692,2699],{"data":1539,"content":1540,"nodeType":254},{},[1541],{"data":1542,"marks":1543,"value":1544,"nodeType":247},{},[],"From massive breaches like the Snowflake incident to novel phishing techniques documented by Push researchers, 2024 was the year that identity attacks left their mark. Looking back over what we saw in the wild and what we found through Push’s own research, three key themes stand out:",{"data":1546,"content":1547,"nodeType":467},{},[1548,1567,1577],{"data":1549,"content":1550,"nodeType":471},{},[1551],{"data":1552,"content":1553,"nodeType":254},{},[1554,1558,1563],{"data":1555,"marks":1556,"value":1557,"nodeType":247},{},[],"Account takeover techniques on cloud apps are fundamentally different from traditional network-based attacks. To have the best chance of preventing account takeover, defenders need to  disrupt attacks ",{"data":1559,"marks":1560,"value":1562,"nodeType":247},{},[1561],{"type":752},"before",{"data":1564,"marks":1565,"value":1566,"nodeType":247},{},[]," they’re successful.",{"data":1568,"content":1569,"nodeType":471},{},[1570],{"data":1571,"content":1572,"nodeType":254},{},[1573],{"data":1574,"marks":1575,"value":1576,"nodeType":247},{},[],"It’s not easy or practical to maintain 100 percent compliance on identity posture standards in a world where employees are using and signing up to apps outside of IT oversight — but it is possible to make this work a lot easier by using tools that help you scale your remediation activities.",{"data":1578,"content":1579,"nodeType":471},{},[1580],{"data":1581,"content":1582,"nodeType":254},{},[1583,1587,1596],{"data":1584,"marks":1585,"value":1586,"nodeType":247},{},[],"Despite another year where cybersecurity spend increased (now up to almost $1,100 per user, according to ",{"data":1588,"content":1590,"nodeType":261},{"uri":1589},"https://www.forrester.com/report/2024-cybersecurity-benchmarks-global/RES181118",[1591],{"data":1592,"marks":1593,"value":1595,"nodeType":247},{},[1594],{"type":269},"Forrester",{"data":1597,"marks":1598,"value":1599,"nodeType":247},{},[],"), existing approaches are not successfully preventing account takeovers. Security teams need to be able to detect and respond to these attacks where they happen: The browser.",{"data":1601,"content":1602,"nodeType":254},{},[1603],{"data":1604,"marks":1605,"value":1606,"nodeType":247},{},[],"In this article, we’ll take a look back at how these themes influenced key features we delivered for Push customers in 2024.",{"data":1608,"content":1609,"nodeType":316},{},[],{"data":1611,"content":1612,"nodeType":243},{},[1613],{"data":1614,"marks":1615,"value":1616,"nodeType":247},{},[],"Defending against modern phishing attacks",{"data":1618,"content":1619,"nodeType":254},{},[1620],{"data":1621,"marks":1622,"value":1623,"nodeType":247},{},[],"Phishing techniques that bypass MFA are now the norm, and few organizations have successfully achieved full coverage of phishing-resistant MFA methods. ",{"data":1625,"content":1626,"nodeType":254},{},[1627,1631,1640,1644,1653],{"data":1628,"marks":1629,"value":1630,"nodeType":247},{},[],"Equally, while phishing attacks via email remain the most commonly reported vector, phishing attacks increasingly target users outside of email. For example, phishing links are often encountered through normal internet use — such as ",{"data":1632,"content":1634,"nodeType":261},{"uri":1633},"https://www.bleepingcomputer.com/news/security/hackers-use-google-search-ads-to-steal-google-ads-accounts/",[1635],{"data":1636,"marks":1637,"value":1639,"nodeType":247},{},[1638],{"type":269},"in malicious Google ads",{"data":1641,"marks":1642,"value":1643,"nodeType":247},{},[]," — and attackers frequently conduct their campaigns over IM platforms like Slack and Teams. Late last year there was ",{"data":1645,"content":1647,"nodeType":261},{"uri":1646},"https://www.linkedin.com/posts/kevin-beaumont-security_ive-been-assisting-a-few-orgs-hit-with-successful-activity-7268055739116445701-xxjZ?utm_source=share&utm_medium=member_desktop",[1648],{"data":1649,"marks":1650,"value":1652,"nodeType":247},{},[1651],{"type":269},"a rise in attackers inundating users with spam via Teams",{"data":1654,"marks":1655,"value":1656,"nodeType":247},{},[],", combined with phone scams posing as IT admins. Since anti-phishing controls are usually email-based, they fail to protect users from attacks taking place elsewhere. ",{"data":1658,"content":1659,"nodeType":254},{},[1660,1664],{"data":1661,"marks":1662,"value":1663,"nodeType":247},{},[],"At Push, we’ve built a suite of anti-phishing features over the last year that act as a defense-in-depth approach to the types of modern phishing techniques we’ve been observing in the wild. ",{"data":1665,"marks":1666,"value":1668,"nodeType":247},{},[1667],{"type":251},"Here’s what we built and why.",{"data":1670,"content":1671,"nodeType":316},{},[],{"data":1673,"content":1674,"nodeType":243},{},[1675],{"data":1676,"marks":1677,"value":1678,"nodeType":247},{},[],"Protecting passwords used for SSO",{"data":1680,"content":1681,"nodeType":335},{},[1682,1687],{"data":1683,"marks":1684,"value":1686,"nodeType":247},{},[1685],{"type":251},"What happened?",{"data":1688,"marks":1689,"value":980,"nodeType":247},{},[],{"data":1691,"content":1692,"nodeType":254},{},[1693],{"data":1694,"marks":1695,"value":1696,"nodeType":247},{},[],"Attackers explicitly targeted Okta, Entra, and Google Workspace accounts in 2023 and 2024, so we knew a top priority would be protecting identity provider accounts. These IdP accounts are a key target because they allow attackers to move laterally to other valuable apps and data via SSO following the initial account takeover.",{"data":1698,"content":1699,"nodeType":254},{},[1700,1704,1713],{"data":1701,"marks":1702,"value":1703,"nodeType":247},{},[],"It’s not just the typical IdPs you need to watch out for, either: Apps like GitHub, Slack, Salesforce, Facebook, X, and others all provide SSO functionality, increasing the blast radius of a compromise. And as we reported in ",{"data":1705,"content":1707,"nodeType":261},{"uri":1706},"https://pushsecurity.com/blog/cross-idp-impersonation/",[1708],{"data":1709,"marks":1710,"value":1712,"nodeType":247},{},[1711],{"type":269},"our research on cross-IdP impersonation",{"data":1714,"marks":1715,"value":1716,"nodeType":247},{},[],", apps can be accessed using multiple SSO methods simultaneously — and 3 in 5 apps that we tested recently did not require re-verification by default when adding a new login method.",{"data":1718,"content":1722,"nodeType":681},{"target":1719},{"sys":1720},{"id":1721,"type":686,"linkType":687},"3EOOr4dVQoiPjl2ucUs1mA",[],{"data":1724,"content":1725,"nodeType":254},{},[1726,1730,1739],{"data":1727,"marks":1728,"value":1729,"nodeType":247},{},[],"Phishing is a problem that would be significantly reduced in a world without passwords. But while the ideal case is that organizations can put in place phishing-resistant authentication methods like passkeys or other WebAuthn-based methods, the reality is that ",{"data":1731,"content":1733,"nodeType":261},{"uri":1732},"https://www.ncsc.gov.uk/blog-post/passkeys-not-perfect-getting-better",[1734],{"data":1735,"marks":1736,"value":1738,"nodeType":247},{},[1737],{"type":269},"it’s not a perfect solution right now",{"data":1740,"marks":1741,"value":1742,"nodeType":247},{},[]," — widespread passkey implementation is hard to achieve.",{"data":1744,"content":1745,"nodeType":254},{},[1746],{"data":1747,"marks":1748,"value":1749,"nodeType":247},{},[],"One of the key advantages of passkeys is that they are domain-bound: Meaning they can’t be used on a site with the wrong domain. So, we started thinking: What if it were possible to essentially domain-bind a password? ",{"data":1751,"content":1752,"nodeType":335},{},[1753,1758],{"data":1754,"marks":1755,"value":1757,"nodeType":247},{},[1756],{"type":251},"What we built",{"data":1759,"marks":1760,"value":980,"nodeType":247},{},[],{"data":1762,"content":1763,"nodeType":254},{},[1764,1768,1778],{"data":1765,"marks":1766,"value":1767,"nodeType":247},{},[],"In the first half of 2024, we delivered our ",{"data":1769,"content":1771,"nodeType":261},{"uri":1770},"https://pushsecurity.com/blog/introducing-sso-password-protection/",[1772],{"data":1773,"marks":1774,"value":1777,"nodeType":247},{},[1775,1776],{"type":269},{"type":251},"SSO password protection",{"data":1779,"marks":1780,"value":1781,"nodeType":247},{},[]," feature, which allows Push administrators to block employees from entering their IdP password into any site that’s not the identity provider — in effect domain-binding SSO credentials. ",{"data":1783,"content":1784,"nodeType":254},{},[1785,1789,1798],{"data":1786,"marks":1787,"value":1788,"nodeType":247},{},[],"Push accomplishes this via the Push browser agent, which ",{"data":1790,"content":1792,"nodeType":261},{"uri":1791},"https://pushsecurity.com/help/10109/#how-does-sso-password-protection-work",[1793],{"data":1794,"marks":1795,"value":1797,"nodeType":247},{},[1796],{"type":269},"observes and fingerprints",{"data":1799,"marks":1800,"value":1801,"nodeType":247},{},[]," the user’s SSO password and legitimate SSO login pages, and then enforces in-browser controls to prevent an SSO password from being submitted on any URL that doesn’t match the legitimate provider, an extremely strong anti-phishing protection. Separately, Push also verifies that passwords it observes are not easily guessable.",{"data":1803,"content":1804,"nodeType":254},{},[1805],{"data":1806,"marks":1807,"value":1808,"nodeType":247},{},[],"The idea behind this approach is to gain some similar benefits to passkeys — by ensuring that passwords used for SSO access to your apps cannot be phished and are unique and strong — but in a way that “just works” with existing password-based authentication. ",{"data":1810,"content":1811,"nodeType":254},{},[1812,1816,1824],{"data":1813,"marks":1814,"value":1815,"nodeType":247},{},[],"Organizations that monitor for SSO password reuse will find that the practice turns out to be incredibly widespread, so being able to detect and prevent password reuse — even outside of actual phishing attempts — is an asset to security teams. (Our ",{"data":1817,"content":1818,"nodeType":261},{"uri":533},[1819],{"data":1820,"marks":1821,"value":1823,"nodeType":247},{},[1822],{"type":269},"research shows",{"data":1825,"marks":1826,"value":1827,"nodeType":247},{},[]," that 10% of IdP accounts are using a password that is shared with another app — where it is much more likely to be compromised.) ",{"data":1829,"content":1833,"nodeType":681},{"target":1830},{"sys":1831},{"id":1832,"type":686,"linkType":687},"4Ce999wf4mqCZwu1jLofsx",[],{"data":1835,"content":1836,"nodeType":254},{},[1837,1841,1850],{"data":1838,"marks":1839,"value":1840,"nodeType":247},{},[],"By streaming events to your SIEM and setting up a simple automation, you can also use Push-supplied intelligence on SSO password reuse to ",{"data":1842,"content":1844,"nodeType":261},{"uri":1843},"https://pushsecurity.com/blog/automating-sso-password-resets-using-push/",[1845],{"data":1846,"marks":1847,"value":1849,"nodeType":247},{},[1848],{"type":269},"automatically reset",{"data":1851,"marks":1852,"value":1853,"nodeType":247},{},[]," potentially compromised passwords — this provides instant response to successful phishing and gets rid of password re-use of your most sensitive credentials in one move - the kind of combo we love!",{"data":1855,"content":1856,"nodeType":316},{},[],{"data":1858,"content":1859,"nodeType":243},{},[1860],{"data":1861,"marks":1862,"value":1863,"nodeType":247},{},[],"Blocking AitM phishing and cloned login pages",{"data":1865,"content":1866,"nodeType":335},{},[1867],{"data":1868,"marks":1869,"value":1686,"nodeType":247},{},[1870],{"type":251},{"data":1872,"content":1873,"nodeType":254},{},[1874],{"data":1875,"marks":1876,"value":1877,"nodeType":247},{},[],"When you’re able to detect SSO passwords being used in all the wrong places, it’s not surprising that one of the main offenders is phishing attacks. ",{"data":1879,"content":1880,"nodeType":254},{},[1881,1885,1894],{"data":1882,"marks":1883,"value":1884,"nodeType":247},{},[],"In 2024, we wrote extensively about the rise in ",{"data":1886,"content":1888,"nodeType":261},{"uri":1887},"https://pushsecurity.com/blog/phishing-2-0-how-phishing-toolkits-are-evolving-with-aitm/",[1889],{"data":1890,"marks":1891,"value":1893,"nodeType":247},{},[1892],{"type":269},"modern phishing attacks",{"data":1895,"marks":1896,"value":1897,"nodeType":247},{},[]," that use adversary-in-the middle toolkits (AiTM), including EvilNoVNC, Evilginx, and others.",{"data":1899,"content":1900,"nodeType":254},{},[1901,1905,1914],{"data":1902,"marks":1903,"value":1904,"nodeType":247},{},[],"AiTM phishing is a newer variant of phishing that allows attackers to bypass MFA protection by using tools that act as a proxy between the end-user and a legitimate login portal. AitM attacks increased 146% in 2023 (",{"data":1906,"content":1908,"nodeType":261},{"uri":1907},"https://cdn-dynmedia-1.microsoft.com/is/content/microsoftcorp/microsoft/final/en-us/microsoft-brand/documents/Microsoft%20Digital%20Defense%20Report%202024%20%281%29.pdf",[1909],{"data":1910,"marks":1911,"value":1913,"nodeType":247},{},[1912],{"type":269},"Microsoft",{"data":1915,"marks":1916,"value":1917,"nodeType":247},{},[],").",{"data":1919,"content":1920,"nodeType":254},{},[1921],{"data":1922,"marks":1923,"value":1924,"nodeType":247},{},[],"This trend in tradecraft was reflected in our own customer base last year, but what’s interesting is that we observed a lot of phish kits and tactics that were new — meaning traditional detections failed to find them before Push did. ",{"data":1926,"content":1927,"nodeType":254},{},[1928,1932,1941],{"data":1929,"marks":1930,"value":1931,"nodeType":247},{},[],"In particular, we saw newer ",{"data":1933,"content":1935,"nodeType":261},{"uri":1934},"https://pushsecurity.com/blog/how-aitm-phishing-kits-evade-detection/",[1936],{"data":1937,"marks":1938,"value":1940,"nodeType":247},{},[1939],{"type":269},"web-based obfuscation techniques",{"data":1942,"marks":1943,"value":1944,"nodeType":247},{},[]," that allowed attackers to get past the features of email security tools like web gateways and email scanning appliances, such as bypassing web sandbox analysis, and deter other forms of automated investigation by using Cloudflare Turnstile and other tactics — similar to the approaches legit websites use to protect against automated bots (this is essentially the same problem for both).",{"data":1946,"content":1947,"nodeType":254},{},[1948],{"data":1949,"marks":1950,"value":1951,"nodeType":247},{},[],"The gap in existing controls was obvious: When all phishing routes eventually lead to the browser, security teams need to be able to detect and respond in the browser. To do this well they need to observe what the employee sees, not what loads in a sandbox.",{"data":1953,"content":1954,"nodeType":335},{},[1955],{"data":1956,"marks":1957,"value":1757,"nodeType":247},{},[1958],{"type":251},{"data":1960,"content":1961,"nodeType":254},{},[1962,1966,1976,1981],{"data":1963,"marks":1964,"value":1965,"nodeType":247},{},[],"To address this gap, we released new capabilities for the Push browser agent to be able to ",{"data":1967,"content":1969,"nodeType":261},{"uri":1968},"https://pushsecurity.com/blog/introducing-aitm-phishing-toolkit-detection-powered-by-the-push-browser/",[1970],{"data":1971,"marks":1972,"value":1975,"nodeType":247},{},[1973,1974],{"type":269},{"type":251},"detect and block",{"data":1977,"marks":1978,"value":1980,"nodeType":247},{},[1979],{"type":251}," when a site is running AiTM phishing toolkits",{"data":1982,"marks":1983,"value":1084,"nodeType":247},{},[],{"data":1985,"content":1986,"nodeType":254},{},[1987],{"data":1988,"marks":1989,"value":1990,"nodeType":247},{},[],"Push does this via a set of readymade detections for common AiTM tools. By dynamically analyzing the behavior of malware in the browser, the Push browser agent can find indicators of compromise beyond just domains, file names, IP addresses, etc., focusing instead on behavioral attributes, such as Javascript calls being made or data structures saved to local storage.",{"data":1992,"content":1993,"nodeType":254},{},[1994,1998,2007],{"data":1995,"marks":1996,"value":1997,"nodeType":247},{},[],"This approach of focusing on the top of the ",{"data":1999,"content":2001,"nodeType":261},{"uri":2000},"https://pushsecurity.com/blog/our-design-philosophy-detecting-what-matters/#id-building-effective-identity-threat-detection-controls_id-scenario-detecting-a-web-based-phishing-attack",[2002],{"data":2003,"marks":2004,"value":2006,"nodeType":247},{},[2005],{"type":269},"Pyramid of Pain",{"data":2008,"marks":2009,"value":2010,"nodeType":247},{},[]," — e.g. building detections for attributes of an attack that are the hardest for attackers to change, and therefore the most reliably accurate — is core to Push’s design philosophy. ",{"data":2012,"content":2013,"nodeType":254},{},[2014,2018,2028],{"data":2015,"marks":2016,"value":2017,"nodeType":247},{},[],"Finally, toward the second half of the year, we released ",{"data":2019,"content":2021,"nodeType":261},{"uri":2020},"https://pushsecurity.com/blog/introducing-cloned-login-page-detection/",[2022],{"data":2023,"marks":2024,"value":2027,"nodeType":247},{},[2025,2026],{"type":269},{"type":251},"cloned login page detection",{"data":2029,"marks":2030,"value":2031,"nodeType":247},{},[],", a natural extension of our layered approach to preventing phishing attacks in the browser. With this security control, you can identify malicious webpages that are masquerading as legitimate IdP login portals. ",{"data":2033,"content":2037,"nodeType":681},{"target":2034},{"sys":2035},{"id":2036,"type":686,"linkType":687},"4y25OxesssUk9lzEx12HFa",[],{"data":2039,"content":2040,"nodeType":254},{},[2041],{"data":2042,"marks":2043,"value":2044,"nodeType":247},{},[],"When a cloned login page is detected, you can add the URL to your blocklist in Push and prevent any other employees from being targeted. ",{"data":2046,"content":2047,"nodeType":254},{},[2048,2052,2061],{"data":2049,"marks":2050,"value":2051,"nodeType":247},{},[],"By layering multiple anti-phishing controls that all prevent account takeover, defenders have the best chance at thwarting the ",{"data":2053,"content":2055,"nodeType":261},{"uri":2054},"https://pushsecurity.com/blog/shifting-detection-left-for-more-effective-itdr/",[2056],{"data":2057,"marks":2058,"value":2060,"nodeType":247},{},[2059],{"type":269},"short, fast attack chains",{"data":2062,"marks":2063,"value":2064,"nodeType":247},{},[]," that are emblematic of today’s identity attacks.",{"data":2066,"content":2067,"nodeType":316},{},[],{"data":2069,"content":2070,"nodeType":243},{},[2071],{"data":2072,"marks":2073,"value":2074,"nodeType":247},{},[],"Defending against stolen sessions and stolen credentials",{"data":2076,"content":2077,"nodeType":254},{},[2078,2082,2091],{"data":2079,"marks":2080,"value":2081,"nodeType":247},{},[],"With as little as $10 to buy a stolen password and a little skill, attackers capitalized on the use of stolen credentials last year. Stolen creds were the No. 1 attacker action in 2023 and 2024, according to ",{"data":2083,"content":2085,"nodeType":261},{"uri":2084},"https://www.verizon.com/business/en-gb/resources/reports/dbir/",[2086],{"data":2087,"marks":2088,"value":2090,"nodeType":247},{},[2089],{"type":269},"Verizon",{"data":2092,"marks":2093,"value":822,"nodeType":247},{},[],{"data":2095,"content":2096,"nodeType":254},{},[2097,2101,2109],{"data":2098,"marks":2099,"value":2100,"nodeType":247},{},[],"Nowhere was this more plain than in the ",{"data":2102,"content":2103,"nodeType":261},{"uri":297},[2104],{"data":2105,"marks":2106,"value":2108,"nodeType":247},{},[2107],{"type":269},"attacks on Snowflake customers",{"data":2110,"marks":2111,"value":2112,"nodeType":247},{},[],", one of the biggest breaches of last year. In this incident, cyber criminals targeted around 165 customers of the cloud-based data warehouse tool Snowflake by taking over accounts using credentials harvested from infostealer infections dating as far back as 2020.",{"data":2114,"content":2115,"nodeType":254},{},[2116],{"data":2117,"marks":2118,"value":2119,"nodeType":247},{},[],"The Snowflake incident underscored the challenges of control and visibility that security teams face when attempting to secure identities on a patchwork of managed and unmanaged apps:",{"data":2121,"content":2122,"nodeType":467},{},[2123,2133,2143,2153,2163],{"data":2124,"content":2125,"nodeType":471},{},[2126],{"data":2127,"content":2128,"nodeType":254},{},[2129],{"data":2130,"marks":2131,"value":2132,"nodeType":247},{},[],"Do I know all the workforce accounts my employees use?",{"data":2134,"content":2135,"nodeType":471},{},[2136],{"data":2137,"content":2138,"nodeType":254},{},[2139],{"data":2140,"marks":2141,"value":2142,"nodeType":247},{},[],"Do those accounts have a strong security posture?",{"data":2144,"content":2145,"nodeType":471},{},[2146],{"data":2147,"content":2148,"nodeType":254},{},[2149],{"data":2150,"marks":2151,"value":2152,"nodeType":247},{},[],"Do those accounts use MFA? The most phishing-resistant methods?",{"data":2154,"content":2155,"nodeType":471},{},[2156],{"data":2157,"content":2158,"nodeType":254},{},[2159],{"data":2160,"marks":2161,"value":2162,"nodeType":247},{},[],"Do I have tools to detect, respond, and remediate after an account takeover or breach of a critical software vendor?",{"data":2164,"content":2165,"nodeType":471},{},[2166],{"data":2167,"content":2168,"nodeType":254},{},[2169],{"data":2170,"marks":2171,"value":2172,"nodeType":247},{},[],"Do I know when a session has been stolen, pointing to a device compromised by infostealer malware?",{"data":2174,"content":2175,"nodeType":254},{},[2176],{"data":2177,"marks":2178,"value":2179,"nodeType":247},{},[],"Here’s what we delivered last year to make it easier for security teams to protect their organizations from the threat of stolen sessions and stolen creds.",{"data":2181,"content":2182,"nodeType":316},{},[],{"data":2184,"content":2185,"nodeType":243},{},[2186],{"data":2187,"marks":2188,"value":2189,"nodeType":247},{},[],"Detecting stolen sessions",{"data":2191,"content":2192,"nodeType":335},{},[2193],{"data":2194,"marks":2195,"value":1686,"nodeType":247},{},[2196],{"type":251},{"data":2198,"content":2199,"nodeType":254},{},[2200,2203,2212],{"data":2201,"marks":2202,"value":29,"nodeType":247},{},[],{"data":2204,"content":2206,"nodeType":261},{"uri":2205},"https://pushsecurity.com/blog/what-the-rise-of-infostealers-says-about-identity-attacks/",[2207],{"data":2208,"marks":2209,"value":2211,"nodeType":247},{},[2210],{"type":269},"Infostealer malware",{"data":2213,"marks":2214,"value":2215,"nodeType":247},{},[]," — a type of malware designed to collect user credentials, including session cookies, from end-user devices — had a very successful 2024, accounting for nearly 10 percent of activity that Red Canary was able to associate with named threats, and the majority of all detected malware that Sophos threat researchers documented last year.",{"data":2217,"content":2218,"nodeType":254},{},[2219,2223,2232],{"data":2220,"marks":2221,"value":2222,"nodeType":247},{},[],"While the use of stolen credentials is rampant, often facilitated by successful infostealer campaigns, a related attack type also ",{"data":2224,"content":2226,"nodeType":261},{"uri":2225},"https://pushsecurity.com/blog/what-the-rise-of-infostealers-says-about-identity-attacks/#id-the-state-of-infostealers-today",[2227],{"data":2228,"marks":2229,"value":2231,"nodeType":247},{},[2230],{"type":269},"jumped in prevalence",{"data":2233,"marks":2234,"value":2235,"nodeType":247},{},[]," last year: session token theft attacks.",{"data":2237,"content":2238,"nodeType":254},{},[2239],{"data":2240,"marks":2241,"value":2242,"nodeType":247},{},[],"Using stolen tokens, adversaries don’t need to bypass MFA directly. They can simply import the tokens into their browser and assume an already authorized session. ",{"data":2244,"content":2245,"nodeType":335},{},[2246],{"data":2247,"marks":2248,"value":1757,"nodeType":247},{},[2249],{"type":251},{"data":2251,"content":2252,"nodeType":254},{},[2253],{"data":2254,"marks":2255,"value":2256,"nodeType":247},{},[],"In order to detect a stolen session in use, you need telemetry that allows you to tie activity to a trusted endpoint. This didn’t previously exist, and you have to be in the browser to do it. So that’s what we built. ",{"data":2258,"content":2259,"nodeType":254},{},[2260,2264,2274],{"data":2261,"marks":2262,"value":2263,"nodeType":247},{},[],"Push’s ",{"data":2265,"content":2267,"nodeType":261},{"uri":2266},"https://pushsecurity.com/blog/introducing-session-token-theft-detection-why-browser-is-best/",[2268],{"data":2269,"marks":2270,"value":2273,"nodeType":247},{},[2271,2272],{"type":269},{"type":251},"session theft detection",{"data":2275,"marks":2276,"value":2277,"nodeType":247},{},[]," capability uses the power of the Push browser extension to inject a unique marker into the user-agent string of sessions that occur in browsers enrolled in Push. ",{"data":2279,"content":2280,"nodeType":254},{},[2281],{"data":2282,"marks":2283,"value":2284,"nodeType":247},{},[],"By analyzing logs from your IdP in your SIEM, you can then identify activity from the same session that both has and that lacks the Push marker, indicating that a session has been extracted from the browser and maliciously imported into a different browser that is not enrolled in Push.",{"data":2286,"content":2287,"nodeType":254},{},[2288],{"data":2289,"marks":2290,"value":2291,"nodeType":247},{},[],"This is a reliable signal that a stolen session token is being used and an endpoint has been compromised.",{"data":2293,"content":2297,"nodeType":681},{"target":2294},{"sys":2295},{"id":2296,"type":686,"linkType":687},"1XNNkaoW64t3PPvC54KGXF",[],{"data":2299,"content":2303,"nodeType":681},{"target":2300},{"sys":2301},{"id":2302,"type":686,"linkType":687},"6dOEnPzZXd9DqeSdalqlzO",[],{"data":2305,"content":2306,"nodeType":316},{},[],{"data":2308,"content":2309,"nodeType":243},{},[2310],{"data":2311,"marks":2312,"value":2313,"nodeType":247},{},[],"Detecting compromised credentials",{"data":2315,"content":2316,"nodeType":335},{},[2317],{"data":2318,"marks":2319,"value":1686,"nodeType":247},{},[2320],{"type":251},{"data":2322,"content":2323,"nodeType":254},{},[2324],{"data":2325,"marks":2326,"value":2327,"nodeType":247},{},[],"Alongside stolen session cookies, stolen credentials made a lot of headlines last year. The 2024 Verizon DBIR found that 79% of web application compromises were the result of breached creds, and researchers at IBM found a 71% year-over-year increase in cyberattacks using stolen or compromised credentials.",{"data":2329,"content":2330,"nodeType":254},{},[2331,2335,2344],{"data":2332,"marks":2333,"value":2334,"nodeType":247},{},[],"In Push’s own research, we counted ",{"data":2336,"content":2338,"nodeType":261},{"uri":2337},"https://pushsecurity.com/blog/2024-identity-breaches/",[2339],{"data":2340,"marks":2341,"value":2343,"nodeType":247},{},[2342],{"type":269},"30 public identity-related breaches",{"data":2345,"marks":2346,"value":2347,"nodeType":247},{},[]," in 2024 where the breach and the breach vector were disclosed. Of those, nearly three-quarters were the result of compromised credentials, including notable breaches such as Microsoft, Change Healthcare, and the attacks on Snowflake customers.",{"data":2349,"content":2350,"nodeType":2358},{},[2351],{"data":2352,"content":2353,"nodeType":254},{},[2354],{"data":2355,"marks":2356,"value":2357,"nodeType":247},{},[],"73% of public identity-related breaches in 2024 were the result of compromised credentials (the rest were phishing attacks). ","blockquote",{"data":2360,"content":2361,"nodeType":254},{},[2362,2366,2374],{"data":2363,"marks":2364,"value":2365,"nodeType":247},{},[],"The influx of compromised credentials has been amplified by the ",{"data":2367,"content":2368,"nodeType":261},{"uri":2205},[2369],{"data":2370,"marks":2371,"value":2373,"nodeType":247},{},[2372],{"type":269},"rise of infostealers",{"data":2375,"marks":2376,"value":2377,"nodeType":247},{},[],", which contribute the vast majority of valid stolen credentials, alongside mass credential phishing campaigns and third-party data breach dumps. ",{"data":2379,"content":2380,"nodeType":254},{},[2381],{"data":2382,"marks":2383,"value":2384,"nodeType":247},{},[],"And while there’s no shortage of threat intelligence about stolen credentials for sale on the web, security teams struggle to separate the needle from the haystack because a large portion of TI on stolen creds is out of date.",{"data":2386,"content":2387,"nodeType":254},{},[2388],{"data":2389,"marks":2390,"value":2391,"nodeType":247},{},[],"In evaluating TI data here at Push, we reviewed 5,763 username and password combos that matched domains in use by Push customers. We found that less than 1% of the creds in a multi-vendor dataset were true positives. In other words, 99.5% of the stolen creds we checked were false positives at the time of review — illustrating the challenge security teams face when trying to extract actionable intelligence from this kind of data. ",{"data":2393,"content":2394,"nodeType":2358},{},[2395],{"data":2396,"content":2397,"nodeType":254},{},[2398],{"data":2399,"marks":2400,"value":2401,"nodeType":247},{},[],"99.5% of the findings in compromised credential feeds were found to be false positives.",{"data":2403,"content":2404,"nodeType":335},{},[2405],{"data":2406,"marks":2407,"value":1757,"nodeType":247},{},[2408],{"type":251},{"data":2410,"content":2411,"nodeType":254},{},[2412],{"data":2413,"marks":2414,"value":2415,"nodeType":247},{},[],"Using its browser agent, Push assesses the strength of end-user passwords by creating and analyzing a truncated, salted SHA256 hash of the password for a given account. (These k-anonymized fingerprints are never seen by Push’s back-end and exist only in local browser extension storage.) ",{"data":2417,"content":2418,"nodeType":254},{},[2419,2423,2432],{"data":2420,"marks":2421,"value":2422,"nodeType":247},{},[],"These fingerprints give Push a directly observable source of truth for corporate creds, which allowed us to build a ",{"data":2424,"content":2425,"nodeType":261},{"uri":424},[2426],{"data":2427,"marks":2428,"value":2431,"nodeType":247},{},[2429,2430],{"type":269},{"type":251},"verified stolen credential detection",{"data":2433,"marks":2434,"value":2435,"nodeType":247},{},[]," capability last year that removes all false positives from TI sources to pinpoint only those stolen creds still actively in use by employees.",{"data":2437,"content":2441,"nodeType":681},{"target":2438},{"sys":2439},{"id":2440,"type":686,"linkType":687},"3BITHZvDadjHpOAqIn0g4w",[],{"data":2443,"content":2444,"nodeType":316},{},[],{"data":2446,"content":2447,"nodeType":243},{},[2448],{"data":2449,"marks":2450,"value":2451,"nodeType":247},{},[],"Reducing and securing shadow IT and account sprawl",{"data":2453,"content":2454,"nodeType":254},{},[2455],{"data":2456,"marks":2457,"value":2458,"nodeType":247},{},[],"You can think of this last part of the story as the ground from which the attack trends we’ve been talking about emerged: The shift to doing business almost entirely in the browser, and the resulting sprawl in accounts and unmanaged apps, leading to an explosion of internet-facing identities for threat actors to target.",{"data":2460,"content":2461,"nodeType":254},{},[2462],{"data":2463,"marks":2464,"value":2465,"nodeType":247},{},[],"Even in organizations with mature security practices, the challenge of getting 100% compliance with identity posture best practices is evident. Last year, Push researchers analyzed a data set of 300,000 accounts from our customer base and found that:",{"data":2467,"content":2468,"nodeType":467},{},[2469,2488,2507],{"data":2470,"content":2471,"nodeType":471},{},[2472],{"data":2473,"content":2474,"nodeType":254},{},[2475,2479,2484],{"data":2476,"marks":2477,"value":2478,"nodeType":247},{},[],"Organizations have ",{"data":2480,"marks":2481,"value":2483,"nodeType":247},{},[2482],{"type":251},"more apps and identities than they thought",{"data":2485,"marks":2486,"value":2487,"nodeType":247},{},[]," — an average of ~15 identities per employee and ~220 apps per organization.",{"data":2489,"content":2490,"nodeType":471},{},[2491],{"data":2492,"content":2493,"nodeType":254},{},[2494,2498,2503],{"data":2495,"marks":2496,"value":2497,"nodeType":247},{},[],"Many accounts ",{"data":2499,"marks":2500,"value":2502,"nodeType":247},{},[2501],{"type":251},"lack basic security protections",{"data":2504,"marks":2505,"value":2506,"nodeType":247},{},[],", with 37% of accounts lacking any form of MFA and ~9% of accounts using a password that is leaked, weak, or reused, making them especially susceptible to account takeover. On accounts where password is the only login method in use (e.g. not using SSO or any other federated login like OIDC), there was no MFA in use in 4 out of 5 cases.",{"data":2508,"content":2509,"nodeType":471},{},[2510],{"data":2511,"content":2512,"nodeType":254},{},[2513,2517,2522,2526,2535],{"data":2514,"marks":2515,"value":2516,"nodeType":247},{},[],"Security ",{"data":2518,"marks":2519,"value":2521,"nodeType":247},{},[2520],{"type":251},"gaps persist even with SSO",{"data":2523,"marks":2524,"value":2525,"nodeType":247},{},[]," accounts — with 10% of SSO-using accounts also having a local password, a risk for ",{"data":2527,"content":2529,"nodeType":261},{"uri":2528},"https://pushsecurity.com/blog/ghost-logins-when-forgotten-identities-come-back-to-haunt-you/",[2530],{"data":2531,"marks":2532,"value":2534,"nodeType":247},{},[2533],{"type":269},"ghost logins",{"data":2536,"marks":2537,"value":2538,"nodeType":247},{},[],"; and 1 in 5 IdP accounts themselves missing MFA.",{"data":2540,"content":2541,"nodeType":254},{},[2542],{"data":2543,"marks":2544,"value":2545,"nodeType":247},{},[],"From our perspective, organizations need scalable controls, and they need easy-to-deploy tools that get them visibility of all their workforce identities, apps, and accounts alongside telemetry that makes the information actionable.",{"data":2547,"content":2548,"nodeType":254},{},[2549],{"data":2550,"marks":2551,"value":2552,"nodeType":247},{},[],"Push already provides a real-time inventory of all your accounts and apps, including internal corporate apps, and analyzes the security posture, login methods, and MFA status of those accounts to offer a comprehensive picture of your identity attack surface. ",{"data":2554,"content":2555,"nodeType":254},{},[2556],{"data":2557,"marks":2558,"value":2559,"nodeType":247},{},[],"To help customers enforce their security policies even more seamlessly, here’s what we built last year:",{"data":2561,"content":2562,"nodeType":335},{},[2563],{"data":2564,"marks":2565,"value":2567,"nodeType":247},{},[2566],{"type":251},"1. App banners",{"data":2569,"content":2570,"nodeType":254},{},[2571,2575,2585],{"data":2572,"marks":2573,"value":2574,"nodeType":247},{},[],"With a range of modes from informing to blocking, ",{"data":2576,"content":2578,"nodeType":261},{"uri":2577},"https://pushsecurity.com/help/10106#start",[2579],{"data":2580,"marks":2581,"value":2584,"nodeType":247},{},[2582,2583],{"type":269},{"type":251},"app banners",{"data":2586,"marks":2587,"value":2588,"nodeType":247},{},[]," allow security teams to communicate best practices and policies with end-users directly in their browser. It works by displaying a banner with your custom message on the login and signup pages for workplace apps. ",{"data":2590,"content":2591,"nodeType":254},{},[2592],{"data":2593,"marks":2594,"value":2595,"nodeType":247},{},[],"Using configuration rules, you can set conditions for how banner controls get applied. Common use cases include: Restricting use of GenAI software; carving out an exception for admins on a specific app; reminding users to log in with SSO instead of a password, and others. ",{"data":2597,"content":2601,"nodeType":681},{"target":2598},{"sys":2599},{"id":2600,"type":686,"linkType":687},"4RPHmeMLyZmb5V8rXYLtey",[],{"data":2603,"content":2604,"nodeType":335},{},[2605,2610],{"data":2606,"marks":2607,"value":2609,"nodeType":247},{},[2608],{"type":251},"2. Password manager identification",{"data":2611,"marks":2612,"value":980,"nodeType":247},{},[],{"data":2614,"content":2615,"nodeType":254},{},[2616,2620,2630],{"data":2617,"marks":2618,"value":2619,"nodeType":247},{},[],"We also expanded Push’s capability to observe employees’ account security posture by adding an identification of ",{"data":2621,"content":2623,"nodeType":261},{"uri":2622},"https://pushsecurity.com/blog/stop-users-saving-corp-creds-into-personal-password-managers/",[2624],{"data":2625,"marks":2626,"value":2629,"nodeType":247},{},[2627,2628],{"type":269},{"type":251},"which password manager",{"data":2631,"marks":2632,"value":2633,"nodeType":247},{},[]," (if any) they’re using. ",{"data":2635,"content":2636,"nodeType":254},{},[2637],{"data":2638,"marks":2639,"value":2640,"nodeType":247},{},[],"We’ve heard from many security teams that they’re concerned about corporate credentials being stored in unapproved password managers — not to mention the ROI from ensuring employees are all using the corporate password manager you already pay for. This feature helps them achieve both objectives.",{"data":2642,"content":2643,"nodeType":335},{},[2644],{"data":2645,"marks":2646,"value":2648,"nodeType":247},{},[2647],{"type":251},"3. MFA enforcement",{"data":2650,"content":2651,"nodeType":254},{},[2652,2656,2665],{"data":2653,"marks":2654,"value":2655,"nodeType":247},{},[],"Finally, we rounded out 2024 with a new security control called ",{"data":2657,"content":2658,"nodeType":261},{"uri":785},[2659],{"data":2660,"marks":2661,"value":2664,"nodeType":247},{},[2662,2663],{"type":269},{"type":251},"MFA enforcement",{"data":2666,"marks":2667,"value":2668,"nodeType":247},{},[]," that builds on the popular app banners concept by detecting when users lack MFA and then prompting them to register for MFA. ",{"data":2670,"content":2671,"nodeType":254},{},[2672],{"data":2673,"marks":2674,"value":2675,"nodeType":247},{},[],"Admins choose which apps they wish to enforce MFA on, and the Push extension does the rest. ",{"data":2677,"content":2678,"nodeType":254},{},[2679],{"data":2680,"marks":2681,"value":2682,"nodeType":247},{},[],"Security teams we work with are especially eager to use this feature to close MFA coverage gaps on non-SSO and otherwise unmanaged applications.",{"data":2684,"content":2688,"nodeType":681},{"target":2685},{"sys":2686},{"id":2687,"type":686,"linkType":687},"4imhff7SWJi2Gan5iFEs2P",[],{"data":2690,"content":2691,"nodeType":316},{},[],{"data":2693,"content":2694,"nodeType":243},{},[2695],{"data":2696,"marks":2697,"value":2698,"nodeType":247},{},[],"Want to see more?",{"data":2700,"content":2701,"nodeType":254},{},[2702,2706,2712],{"data":2703,"marks":2704,"value":2705,"nodeType":247},{},[],"There’s a lot we didn’t touch on here that Push can help you achieve. If you’d like to learn more, ",{"data":2707,"content":2708,"nodeType":261},{"uri":1499},[2709],{"data":2710,"marks":2711,"value":1502,"nodeType":247},{},[],{"data":2713,"marks":2714,"value":1506,"nodeType":247},{},[],"How real-world attacks and research drove Push’s most popular features of 2024","How in-the-wild attacks and our own R&D inspired what we built in 2024 to stop account takeover and reduce security risks across your workforce identities. ","2025-01-16T00:00:00.000Z","push-features-2024",{"items":2720},[2721,2725],{"sys":2722,"name":2724},{"id":2723},"4ksQNCFeBf8H4QIORqpRLw","Detection & response",{"sys":2726,"name":1522},{"id":1521},{"items":2728},[2729],{"fullName":2730,"firstName":2731,"jobTitle":2732,"profilePicture":2733},"Kelly Davenport","Kelly","Product Team",{"url":2734},"https://images.ctfassets.net/y1cdw1ablpvd/1hi8bEuVfn5sF57LivAq6d/9a3b82426c697d765e2e450e33a18424/kelly_profile_pic.jpeg",{"__typename":886,"sys":2736,"content":2738,"title":3383,"synopsis":3384,"hashTags":62,"publishedDate":2717,"slug":3385,"tagsCollection":3386,"authorsCollection":3392},{"id":2737},"wikyVxlHwKUOKM9xo19eP",{"json":2739},{"nodeType":239,"data":2740,"content":2741},{},[2742,2748,2751,2758,2781,2811,2821,2840,2847,2853,2860,2875,2882,2885,2892,2899,2918,2925,2971,2978,2984,2990,2997,3030,3044,3047,3054,3061,3068,3075,3082,3089,3096,3204,3210,3225,3232,3247,3280,3296,3303,3318,3324,3331,3338,3344,3351,3357,3364],{"nodeType":681,"data":2743,"content":2747},{"target":2744},{"sys":2745},{"id":2746,"type":686,"linkType":687},"1hUpsNwuhEXwSPijvRflTq",[],{"nodeType":316,"data":2749,"content":2750},{},[],{"nodeType":254,"data":2752,"content":2753},{},[2754],{"nodeType":247,"value":2755,"marks":2756,"data":2757},"There are two things every security operations engineer can agree on:",[],{},{"nodeType":467,"data":2759,"content":2760},{},[2761,2771],{"nodeType":471,"data":2762,"content":2763},{},[2764],{"nodeType":254,"data":2765,"content":2766},{},[2767],{"nodeType":247,"value":2768,"marks":2769,"data":2770},"Get MFA on every account on every app.",[],{},{"nodeType":471,"data":2772,"content":2773},{},[2774],{"nodeType":254,"data":2775,"content":2776},{},[2777],{"nodeType":247,"value":2778,"marks":2779,"data":2780},"This is stupidly harder to achieve than it seems.",[],{},{"nodeType":254,"data":2782,"content":2783},{},[2784,2788,2796,2800,2807],{"nodeType":247,"value":2785,"marks":2786,"data":2787},"The penalties for failing to solve this hard simple problem are abundantly clear. Stolen credentials accounted for roughly half of the initial access methods observed this year across 30,000+ attacks, according to Verizon’s 2024 ",[],{},{"nodeType":261,"data":2789,"content":2790},{"uri":263},[2791],{"nodeType":247,"value":2792,"marks":2793,"data":2795},"Data Breach Investigations Report",[2794],{"type":269},{},{"nodeType":247,"value":2797,"marks":2798,"data":2799},". And ",[],{},{"nodeType":261,"data":2801,"content":2802},{"uri":2337},[2803],{"nodeType":247,"value":2804,"marks":2805,"data":2806},"in a review of 30 publicly disclosed breaches involving identity attacks",[],{},{"nodeType":247,"value":2808,"marks":2809,"data":2810}," in 2024, we found that 73% (almost three-quarters) were the result of compromised credentials, with the rest the result of phishing. ",[],{},{"nodeType":2358,"data":2812,"content":2813},{},[2814],{"nodeType":254,"data":2815,"content":2816},{},[2817],{"nodeType":247,"value":2818,"marks":2819,"data":2820},"Three-quarters of publicly disclosed breaches involving identity attacks in 2024 involved compromised credentials and missing MFA.",[],{},{"nodeType":254,"data":2822,"content":2823},{},[2824,2828,2836],{"nodeType":247,"value":2825,"marks":2826,"data":2827},"In the case of the ",[],{},{"nodeType":261,"data":2829,"content":2830},{"uri":297},[2831],{"nodeType":247,"value":2832,"marks":2833,"data":2835},"Snowflake incident",[2834],{"type":269},{},{"nodeType":247,"value":2837,"marks":2838,"data":2839}," earlier this year, a lack of MFA meant the difference between an enormous and murky firefight to clean up accounts breached with legitimate credentials, and a decent night’s sleep. The result was hundreds of millions of breached customer records, nine publicly named victims, and at least one ransom paid.",[],{},{"nodeType":254,"data":2841,"content":2842},{},[2843],{"nodeType":247,"value":2844,"marks":2845,"data":2846},"“Do you know how many accounts we have on this third-party service, who owns them, how many tenants, whether those creds are shared elsewhere, and their security posture?” is not a fun question to answer on a Friday. ",[],{},{"nodeType":681,"data":2848,"content":2852},{"target":2849},{"sys":2850},{"id":2851,"type":686,"linkType":687},"6hg6PLXWMZaEDnGekHEzmD",[],{"nodeType":254,"data":2854,"content":2855},{},[2856],{"nodeType":247,"value":2857,"marks":2858,"data":2859},"For SecOps teams we’ve helped here at Push that responded to incidents affecting third-party apps (like Snowflake), the first item on the recovery plan is to finally solve that hard simple problem: No more MFA gaps.",[],{},{"nodeType":254,"data":2861,"content":2862},{},[2863,2867,2871],{"nodeType":247,"value":2864,"marks":2865,"data":2866},"With our latest feature release, ",[],{},{"nodeType":247,"value":2664,"marks":2868,"data":2870},[2869],{"type":251},{},{"nodeType":247,"value":2872,"marks":2873,"data":2874},", this is so much easier. With MFA enforcement, Push administrators can configure a control to prompt employees to enroll in MFA whenever Push detects that they’re not registered — even on apps that don’t natively provide any administrative enforcement option for MFA. This capability is made possible by the Push browser extension, which uses in-browser messaging and simple workflows to guide users right where they work.",[],{},{"nodeType":254,"data":2876,"content":2877},{},[2878],{"nodeType":247,"value":2879,"marks":2880,"data":2881},"In this article, we’ll cover how Push helps you identify and close MFA gaps, how our new enforcement feature is one part of that solution, and how you can test the platform yourself.",[],{},{"nodeType":316,"data":2883,"content":2884},{},[],{"nodeType":243,"data":2886,"content":2887},{},[2888],{"nodeType":247,"value":2889,"marks":2890,"data":2891},"Shining a light on MFA gaps",[],{},{"nodeType":254,"data":2893,"content":2894},{},[2895],{"nodeType":247,"value":2896,"marks":2897,"data":2898},"There’s no question that the rise of ubiquitous multi-factor authentication has been an enormous advance for defenders in cybersecurity. ",[],{},{"nodeType":254,"data":2900,"content":2901},{},[2902,2906,2915],{"nodeType":247,"value":2903,"marks":2904,"data":2905},"Yet several years into this journey, the problem of verifying and enforcing MFA coverage across an organization remains a bit of a ",[],{},{"nodeType":261,"data":2907,"content":2909},{"uri":2908},"https://en.wikipedia.org/wiki/Puzzle_box",[2910],{"nodeType":247,"value":2911,"marks":2912,"data":2914},"puzzle box",[2913],{"type":269},{},{"nodeType":247,"value":822,"marks":2916,"data":2917},[],{},{"nodeType":254,"data":2919,"content":2920},{},[2921],{"nodeType":247,"value":2922,"marks":2923,"data":2924},"Why is this?",[],{},{"nodeType":467,"data":2926,"content":2927},{},[2928,2938,2948],{"nodeType":471,"data":2929,"content":2930},{},[2931],{"nodeType":254,"data":2932,"content":2933},{},[2934],{"nodeType":247,"value":2935,"marks":2936,"data":2937},"Complex overlapping (and occasionally contradictory) configurations for enterprise MFA solutions can result in entire employee groups not registered for MFA, and other critical missing pieces.",[],{},{"nodeType":471,"data":2939,"content":2940},{},[2941],{"nodeType":254,"data":2942,"content":2943},{},[2944],{"nodeType":247,"value":2945,"marks":2946,"data":2947},"With a sprawling ecosystem of both SSO-managed and unmanaged self-adopted SaaS, MFA coverage ends up looking more like a patchwork than a unified layer of protection. Security teams lack visibility of freemium and self-purchased apps, and when signup is simple, many users will naturally skip MFA registration to remove a layer of friction. The end result is often a suite of core apps managed via SSO that enforce MFA — and a lot of other unmanaged apps that don’t (true nightmare fodder).",[],{},{"nodeType":471,"data":2949,"content":2950},{},[2951],{"nodeType":254,"data":2952,"content":2953},{},[2954,2958,2967],{"nodeType":247,"value":2955,"marks":2956,"data":2957},"Another annoying piece of the puzzle box: Even in organizations with a high adoption rate of phishing-resistant MFA methods, having backup MFA methods (and a lack of total visibility into all of those registered methods) can create situations where ",[],{},{"nodeType":261,"data":2959,"content":2961},{"uri":2960},"https://github.com/pushsecurity/saas-attacks/blob/main/techniques/mfa_downgrade/description.md",[2962],{"nodeType":247,"value":2963,"marks":2964,"data":2966},"MFA downgrade attacks",[2965],{"type":269},{},{"nodeType":247,"value":2968,"marks":2969,"data":2970}," are still possible. In MFA downgrade attacks, backup MFA methods that are less secure such as SMS or TOTP can be exploited, effectively bypassing more phishing-resistant methods.",[],{},{"nodeType":254,"data":2972,"content":2973},{},[2974],{"nodeType":247,"value":2975,"marks":2976,"data":2977},"The challenges of solving this puzzle are evident. ",[],{},{"nodeType":681,"data":2979,"content":2983},{"target":2980},{"sys":2981},{"id":2982,"type":686,"linkType":687},"2BBiFx8pHjSCeLTlP6n6da",[],{"nodeType":681,"data":2985,"content":2989},{"target":2986},{"sys":2987},{"id":2988,"type":686,"linkType":687},"2QnWVpPYRyJQaQ5TuKSSLp",[],{"nodeType":254,"data":2991,"content":2992},{},[2993],{"nodeType":247,"value":2994,"marks":2995,"data":2996},"To shine a light on MFA gaps, then, security teams need three things:",[],{},{"nodeType":467,"data":2998,"content":2999},{},[3000,3010,3020],{"nodeType":471,"data":3001,"content":3002},{},[3003],{"nodeType":254,"data":3004,"content":3005},{},[3006],{"nodeType":247,"value":3007,"marks":3008,"data":3009},"A full accounting of their identity attack surface, including accounts on unmanaged and freemium apps not on SSO.",[],{},{"nodeType":471,"data":3011,"content":3012},{},[3013],{"nodeType":254,"data":3014,"content":3015},{},[3016],{"nodeType":247,"value":3017,"marks":3018,"data":3019},"A trustworthy out-of-band method for verifying MFA coverage, beyond the tangle of conditional access rules.",[],{},{"nodeType":471,"data":3021,"content":3022},{},[3023],{"nodeType":254,"data":3024,"content":3025},{},[3026],{"nodeType":247,"value":3027,"marks":3028,"data":3029},"Visibility into which MFA methods are registered to a given account.",[],{},{"nodeType":254,"data":3031,"content":3032},{},[3033,3037,3041],{"nodeType":247,"value":3034,"marks":3035,"data":3036},"You can get all three with the Push platform. The missing piece we’ve now added is a way to automatically prompt employees to add MFA wherever it’s missing. Enter ",[],{},{"nodeType":247,"value":2664,"marks":3038,"data":3040},[3039],{"type":251},{},{"nodeType":247,"value":822,"marks":3042,"data":3043},[],{},{"nodeType":316,"data":3045,"content":3046},{},[],{"nodeType":243,"data":3048,"content":3049},{},[3050],{"nodeType":247,"value":3051,"marks":3052,"data":3053},"How Push helps you ensure MFA coverage",[],{},{"nodeType":254,"data":3055,"content":3056},{},[3057],{"nodeType":247,"value":3058,"marks":3059,"data":3060},"Let’s take a look at a hypothetical incident response scenario to see how Push’s identity visibility and security controls help you ensure MFA coverage.",[],{},{"nodeType":254,"data":3062,"content":3063},{},[3064],{"nodeType":247,"value":3065,"marks":3066,"data":3067},"We’ll assume that prior to this incident, you had already deployed the Push browser extension, which you can install and enforce using any MDM solution, on all major browsers.",[],{},{"nodeType":254,"data":3069,"content":3070},{},[3071],{"nodeType":247,"value":3072,"marks":3073,"data":3074},"It’s a Friday afternoon (sorry).",[],{},{"nodeType":254,"data":3076,"content":3077},{},[3078],{"nodeType":247,"value":3079,"marks":3080,"data":3081},"News breaks that there’s been a suspected breach at a popular enterprise SaaS service.",[],{},{"nodeType":254,"data":3083,"content":3084},{},[3085],{"nodeType":247,"value":3086,"marks":3087,"data":3088},"You’re familiar with the service, but you don’t believe it’s a core managed app at your organization. Unfortunately, that does not mean you don’t have accounts (sorry again).",[],{},{"nodeType":254,"data":3090,"content":3091},{},[3092],{"nodeType":247,"value":3093,"marks":3094,"data":3095},"Using Push, you can:",[],{},{"nodeType":467,"data":3097,"content":3098},{},[3099,3118,3128,3155,3182],{"nodeType":471,"data":3100,"content":3101},{},[3102],{"nodeType":254,"data":3103,"content":3104},{},[3105,3109,3114],{"nodeType":247,"value":3106,"marks":3107,"data":3108},"Immediately check whether the Push extension has observed employee usage of the breached app. It will appear on the ",[],{},{"nodeType":247,"value":3110,"marks":3111,"data":3113},"Apps",[3112],{"type":251},{},{"nodeType":247,"value":3115,"marks":3116,"data":3117}," table. From this overview, you can see how many accounts Push has seen on that app and how they are accessing it (SSO vs. other methods, such as local password login).",[],{},{"nodeType":471,"data":3119,"content":3120},{},[3121],{"nodeType":254,"data":3122,"content":3123},{},[3124],{"nodeType":247,"value":3125,"marks":3126,"data":3127},"For those accounts on the breached app, you can quickly see whether they have MFA, and which methods are registered. To determine MFA status, the Push extension uses the existing user’s active session on an app to query that account’s MFA registration status using the app’s own API, providing a trustworthy verification. ",[],{},{"nodeType":471,"data":3129,"content":3130},{},[3131],{"nodeType":254,"data":3132,"content":3133},{},[3134,3138,3143,3147,3152],{"nodeType":247,"value":3135,"marks":3136,"data":3137},"You can also see whether the users’ passwords have any security issues, such as a verified stolen credential, or a password that’s weak or reused by filtering the ",[],{},{"nodeType":247,"value":3139,"marks":3140,"data":3142},"Accounts",[3141],{"type":251},{},{"nodeType":247,"value":3144,"marks":3145,"data":3146}," list for ",[],{},{"nodeType":247,"value":3148,"marks":3149,"data":3151},"Findings",[3150],{"type":251},{},{"nodeType":247,"value":822,"marks":3153,"data":3154},[],{},{"nodeType":471,"data":3156,"content":3157},{},[3158],{"nodeType":254,"data":3159,"content":3160},{},[3161,3165,3169,3173,3178],{"nodeType":247,"value":3162,"marks":3163,"data":3164},"For accounts that lack MFA, you can then configure the ",[],{},{"nodeType":247,"value":2664,"marks":3166,"data":3168},[3167],{"type":251},{},{"nodeType":247,"value":3170,"marks":3171,"data":3172}," control from the ",[],{},{"nodeType":247,"value":3174,"marks":3175,"data":3177},"Controls",[3176],{"type":251},{},{"nodeType":247,"value":3179,"marks":3180,"data":3181}," page. This will prompt employees who lack MFA to set it up whenever they next use the app. In parallel, you can reach out to affected employees through your preferred comms channel and ask them to immediately register for MFA and change their password on the app. ",[],{},{"nodeType":471,"data":3183,"content":3184},{},[3185],{"nodeType":254,"data":3186,"content":3187},{},[3188,3192,3201],{"nodeType":247,"value":3189,"marks":3190,"data":3191},"Then use Push’s webhooks to monitor for MFA registrations and password changes to roll in, by querying the ",[],{},{"nodeType":261,"data":3193,"content":3195},{"uri":3194},"https://pushsecurity.redoc.ly/webhooks-v1#operation/login-event",[3196],{"nodeType":247,"value":3197,"marks":3198,"data":3200},"Login event",[3199],{"type":269},{},{"nodeType":247,"value":822,"marks":3202,"data":3203},[],{},{"nodeType":681,"data":3205,"content":3209},{"target":3206},{"sys":3207},{"id":3208,"type":686,"linkType":687},"4OVJU6FRSVU9j1WB9NGyJ4",[],{"nodeType":254,"data":3211,"content":3212},{},[3213,3217,3221],{"nodeType":247,"value":3214,"marks":3215,"data":3216},"By combining visibility of your workforce identities — including granular context on their MFA registration status, MFA methods, and password security, even on unmanaged apps — with in-browser controls like ",[],{},{"nodeType":247,"value":2664,"marks":3218,"data":3220},[3219],{"type":251},{},{"nodeType":247,"value":3222,"marks":3223,"data":3224},", Push helps security teams respond quickly and with assurance that they have the right information and tools to remediate the issue.",[],{},{"nodeType":243,"data":3226,"content":3227},{},[3228],{"nodeType":247,"value":3229,"marks":3230,"data":3231},"A closer look at MFA enforcement",[],{},{"nodeType":254,"data":3233,"content":3234},{},[3235,3239,3243],{"nodeType":247,"value":3236,"marks":3237,"data":3238},"With the in-browser ",[],{},{"nodeType":247,"value":2664,"marks":3240,"data":3242},[3241],{"type":251},{},{"nodeType":247,"value":3244,"marks":3245,"data":3246}," control, we chose this approach to close the loop on missing MFA issues because:",[],{},{"nodeType":467,"data":3248,"content":3249},{},[3250,3260,3270],{"nodeType":471,"data":3251,"content":3252},{},[3253],{"nodeType":254,"data":3254,"content":3255},{},[3256],{"nodeType":247,"value":3257,"marks":3258,"data":3259},"It meets users where they are, in the most relevant context where they can successfully address the issue.",[],{},{"nodeType":471,"data":3261,"content":3262},{},[3263],{"nodeType":254,"data":3264,"content":3265},{},[3266],{"nodeType":247,"value":3267,"marks":3268,"data":3269},"It solves the problem of enforcing MFA on apps that are outside of administrative control — or that don’t provide any administrative controls to enforce MFA registration natively.",[],{},{"nodeType":471,"data":3271,"content":3272},{},[3273],{"nodeType":254,"data":3274,"content":3275},{},[3276],{"nodeType":247,"value":3277,"marks":3278,"data":3279},"It’s tenant-agnostic. That means that you can enforce MFA for a given app on all tenants of that app, even those free-tier or test tenants that you don’t know about and have no control over.",[],{},{"nodeType":254,"data":3281,"content":3282},{},[3283,3287,3292],{"nodeType":247,"value":3284,"marks":3285,"data":3286},"As a happy side effect, your compliance team will thank you for finally allowing them to attest to where MFA is ",[],{},{"nodeType":247,"value":3288,"marks":3289,"data":3291},"actually",[3290],{"type":752},{},{"nodeType":247,"value":3293,"marks":3294,"data":3295}," enforced — with verified results, visible at the account level in Push’s admin reporting — across your environment.",[],{},{"nodeType":254,"data":3297,"content":3298},{},[3299],{"nodeType":247,"value":3300,"marks":3301,"data":3302},"Here’s a closer look at how it works:",[],{},{"nodeType":254,"data":3304,"content":3305},{},[3306,3310,3314],{"nodeType":247,"value":3307,"marks":3308,"data":3309},"To enable MFA enforcement, use the configuration tile on the ",[],{},{"nodeType":247,"value":3174,"marks":3311,"data":3313},[3312],{"type":251},{},{"nodeType":247,"value":3315,"marks":3316,"data":3317}," page of the Push admin console and select which apps should require MFA registration. The control currently works with ~90 high-value apps, including Postman, Retool, Datadog, Atlassian, Okta, and others.",[],{},{"nodeType":681,"data":3319,"content":3323},{"target":3320},{"sys":3321},{"id":3322,"type":686,"linkType":687},"2sDbYZL4oJDxLMbYErJfIN",[],{"nodeType":254,"data":3325,"content":3326},{},[3327],{"nodeType":247,"value":3328,"marks":3329,"data":3330},"You can then customize the message the employees will see.",[],{},{"nodeType":254,"data":3332,"content":3333},{},[3334],{"nodeType":247,"value":3335,"marks":3336,"data":3337},"On the end-user side, employees will see a banner with your message as soon as they use an app where they lack MFA. ",[],{},{"nodeType":681,"data":3339,"content":3343},{"target":3340},{"sys":3341},{"id":3342,"type":686,"linkType":687},"37aH1maXXkF8DxgjUod5dn",[],{"nodeType":254,"data":3345,"content":3346},{},[3347],{"nodeType":247,"value":3348,"marks":3349,"data":3350},"To complete MFA registration, the user can go directly to the app’s MFA registration page from a link in the banner (Push provides this link automatically, where one exists). The extension will query the user’s MFA status regularly in the background and when MFA registration is completed, the banner will disappear and the Push platform will clear the “No MFA” security finding for that account.",[],{},{"nodeType":681,"data":3352,"content":3356},{"target":3353},{"sys":3354},{"id":3355,"type":686,"linkType":687},"3yb4KjhH3AbvvSnfMbNONr",[],{"nodeType":243,"data":3358,"content":3359},{},[3360],{"nodeType":247,"value":3361,"marks":3362,"data":3363},"Find out more",[],{},{"nodeType":254,"data":3365,"content":3366},{},[3367,3371,3379],{"nodeType":247,"value":3368,"marks":3369,"data":3370},"To test our MFA visibility and control features, ",[],{},{"nodeType":261,"data":3372,"content":3374},{"uri":3373},"/demo",[3375],{"nodeType":247,"value":3376,"marks":3377,"data":3378},"request a demo",[],{},{"nodeType":247,"value":3380,"marks":3381,"data":3382}," from our team. We look forward to helping you finally turn the challenge of MFA coverage into a simple problem, easily solved.",[],{},"No more hard simple problems: Enforce MFA on third-party apps with Push","Using Push to enforce MFA on third-party apps in the browser — even where MFA enforcement isn't supported by the app itself.","enforce-mfa-on-third-party-apps",{"items":3387},[3388,3390],{"sys":3389,"name":1522},{"id":1521},{"sys":3391,"name":2724},{"id":2723},{"items":3393},[3394],{"fullName":2730,"firstName":2731,"jobTitle":2732,"profilePicture":3395},{"url":2734},"how-new-ai-agents-will-transform-credential-stuffing-attacks","blog/how-new-ai-agents-will-transform-credential-stuffing-attacks",{"json":3399},{"data":3400,"content":3401,"nodeType":239},{},[3402],{"data":3403,"content":3404,"nodeType":254},{},[3405],{"data":3406,"marks":3407,"value":3408,"nodeType":247},{},[],"Credential stuffing attacks had a huge impact in 2024, fueled by a vicious circle of infostealer infections and data breaches. But things could be about to get worse still with Computer-Using Agents, a new kind of AI agent that enables low-cost, low-effort automation of common web tasks — including those frequently performed by attackers.","Credential stuffing attacks had a huge impact in 2024. But things could be dialled up even further with Computer-Using Agents like OpenAI Operator. ",{"id":3411,"publishedAt":3412},"hpHG3MIyOo5AMxoeLDeBX","2026-01-30T09:13:19.890Z",{"items":3414},[3415,3417],{"sys":3416,"name":1518},{"id":1517},{"sys":3418,"name":1522},{"id":1521},"K0QV56ZX8bGuINzDAWO7BfUPoFgCTML1MQSN__2Ghcg",1784196717556]