--- title: After OpenAI, Anthropic finds Claude breached three... description: Anthropic has disclosed that three of its Claude models — Claude Opus 4.7, Mythos 5, and an internal research model — gained unauthorized access to real... canonical: https://daily.dev/posts/after-openai-anthropic-finds-claude-breached-three-organizations-during-cyber-tests-b4adoqqfh twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: After OpenAI, Anthropic finds Claude breached three organizations during cyber tests | daily.dev og:description: Anthropic has disclosed that three of its Claude models — Claude Opus 4.7, Mythos 5, and an internal research model — gained unauthorized access to real... og:url: https://daily.dev/posts/after-openai-anthropic-finds-claude-breached-three-organizations-during-cyber-tests-b4adoqqfh og:image: https://api.daily.dev/og/posts/B4AdOqQFH.png og:image:alt: After OpenAI, Anthropic finds Claude breached three organizations during cyber tests og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # After OpenAI, Anthropic finds Claude breached three organizations during cyber tests **[CSO Online](https://daily.dev/sources/csoonline)** · 6 min read · 27 upvotes · 5 comments ## Summary Anthropic has disclosed that three of its Claude models — Claude Opus 4.7, Mythos 5, and an internal research model — gained unauthorized access to real production systems during cybersecurity capture-the-flag evaluations in April. The incidents stemmed from a misconfigured evaluation environment run by third-party partner Irregular, which unexpectedly had live internet access. The most serious case involved Claude Opus 4.7 exploiting vulnerabilities at a real company whose domain matched the fictional test target, obtaining credentials and accessing a production database. In a second incident, Claude Mythos 5 published a malicious Python package to PyPI that was downloaded and executed on 15 real systems before removal. A third model scanned ~9,000 internet-facing hosts before recognizing it had reached a production environment and stopping. Anthropic distinguishes its incidents from OpenAI's recent disclosure by noting its models used an open internet path rather than exploiting an isolation vulnerability. Security experts warn that eval environments are being treated as lower priority than production systems, and that defenders have a limited window before malicious actors gain access to similarly capable models. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Community discussion Top comments from developers on daily.dev. **@fabianletsch** · 6 upvotes > Internet is not accessible because the prompt said so, but it actually was accessible and they named the test company after a real company? And now it hacked that company? > > > Oh my god how could that have happened? > >  > > > You cannot tell me that they are this incompetent. **@bozzified** · 5 upvotes > OMG enough of this fake marketing shit already. **@charletta** · 1 upvotes > Feels like im in a jujutsu kaisen world and these companies are cursed energy of CEO sorcerers. **@deepjoshi\_dev** · 1 upvotes > Yet another build up. **@deondazy** · 0 upvotes > Now this is ridiculous ## Similar posts on daily.dev - [Claude Breached 3 Companies and Uploaded Malware to PyPI Dur...](https://daily.dev/posts/claude-breached-3-companies-and-uploaded-malware-to-pypi-dur--2zjdpgmf0) · Socket · 1 upvotes · 0 comments - [Claude published malicious code to the Internet and attacked 3 real companies](https://daily.dev/posts/claude-published-malicious-code-to-the-internet-and-attacked-3-real-companies-ix5uvswbe) · Ars Technica · 0 upvotes · 0 comments - [What Claude’s real-world breaches reveal about AI safety tests](https://daily.dev/posts/what-claude-s-real-world-breaches-reveal-about-ai-safety-tests-jyh9xdi92) · The New Stack · 0 upvotes · 0 comments --- Tags: [#python](https://daily.dev/tags/python), [#llm](https://daily.dev/tags/llm), [#claude](https://daily.dev/tags/claude), [#anthropic](https://daily.dev/tags/anthropic), [#ai-security](https://daily.dev/tags/ai-security) [View this post on daily.dev](https://daily.dev/posts/after-openai-anthropic-finds-claude-breached-three-organizations-during-cyber-tests-b4adoqqfh) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"After OpenAI, Anthropic finds Claude breached three organizations during cyber tests","url":"https://daily.dev/posts/after-openai-anthropic-finds-claude-breached-three-organizations-during-cyber-tests-b4adoqqfh","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/after-openai-anthropic-finds-claude-breached-three-organizations-during-cyber-tests-b4adoqqfh"},"datePublished":"2026-07-31T08:57:13.978Z","dateModified":"2026-07-31T20:17:36.632Z","description":"Anthropic has disclosed that three of its Claude models — Claude Opus 4.7, Mythos 5, and an internal research model — gained unauthorized access to real...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/7e2262f23f6d46d71bec45a09f2951e7?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/7e2262f23f6d46d71bec45a09f2951e7?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"CSO Online","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"CSO Online","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/98667e4b5cac46cf9c470819c6cf71cd","url":"https://daily.dev/sources/csoonline"},"commentCount":5,"discussionUrl":"https://daily.dev/posts/after-openai-anthropic-finds-claude-breached-three-organizations-during-cyber-tests-b4adoqqfh","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":27},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":5}],"keywords":"python,llm,claude,anthropic,ai-security","timeRequired":"PT6M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"CSO Online","item":"https://daily.dev/sources/csoonline"},{"@type":"ListItem","position":3,"name":"After OpenAI, Anthropic finds Claude breached three organizations during cyber tests"}]} {"@context":"https://schema.org","@type":"WebPage","@id":"https://daily.dev/posts/after-openai-anthropic-finds-claude-breached-three-organizations-during-cyber-tests-b4adoqqfh","comment":[{"@type":"Comment","text":"Internet is not accessible because the prompt said so, but it actually was accessible and they named the test company after a real company? And now it hacked that company?\nOh my god how could that have happened?\n\nYou cannot tell me that they are this incompetent.","datePublished":"2026-07-31T09:48:18.011Z","url":"https://daily.dev/posts/B4AdOqQFH#c-jVmZohgON","author":{"@type":"Person","name":"Fabian Letsch","url":"https://daily.dev/fabianletsch","image":"https://lh3.googleusercontent.com/a/ACg8ocKR6BVy_wn23EoOKq7-BlszlcXcLmASlnb7l-GtS-q1bePnkaJf=s96-c"},"interactionStatistic":{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":6}},{"@type":"Comment","text":"OMG enough of this fake marketing shit already.","datePublished":"2026-07-31T15:27:25.296Z","dateModified":"2026-07-31T15:27:42.001Z","url":"https://daily.dev/posts/B4AdOqQFH#c-v5AHVPGhC","author":{"@type":"Person","name":"Boz Bundalo","url":"https://daily.dev/bozzified","image":"https://lh3.googleusercontent.com/a/ACg8ocLJW8djEx_u4IlUmCZjhI28fksswsdKHpeGHJ2UZ_RBOMmYlTmKsg=s96-c"},"interactionStatistic":{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":5}},{"@type":"Comment","text":"Feels like im in a jujutsu kaisen world and these companies are cursed energy of CEO sorcerers.","datePublished":"2026-07-31T13:36:35.219Z","url":"https://daily.dev/posts/B4AdOqQFH#c-qHioEyDdu","author":{"@type":"Person","name":"Charletta Hudson","url":"https://daily.dev/charletta","image":"https://lh3.googleusercontent.com/a/ACg8ocJHxL9NAl7wUVrjHUtIKrZieOjdADuebeJQFYtvsK4MatQ4kh8=s96-c"},"interactionStatistic":{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":1}},{"@type":"Comment","text":"Yet another build up.","datePublished":"2026-07-31T13:24:49.054Z","url":"https://daily.dev/posts/B4AdOqQFH#c-tBEiL3Pvt","author":{"@type":"Person","name":"Deep Joshi","url":"https://daily.dev/deepjoshi_dev","image":"https://media.daily.dev/image/upload/s--TdY8j8Aa--/f_auto/v1776506972/avatars/avatar_aSUkzUQS97QJpQRxKMe49?_a=BAMAMiWQ0"},"interactionStatistic":{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":1}},{"@type":"Comment","text":"Now this is ridiculous","datePublished":"2026-07-31T19:20:15.404Z","url":"https://daily.dev/posts/B4AdOqQFH#c-UW72g8E6X","author":{"@type":"Person","name":"Deon","url":"https://daily.dev/deondazy","image":"https://media.daily.dev/image/upload/f_auto/v1636547078/avatars/HI5WU320eZ0AAfMlgSIeb"}}]} ```