{"slug": "ai-agents-found-paths-into-anthropics-business-systems-in-under-12-hours-startup", "title": "AI agents found paths into Anthropic’s business systems in under 12 hours, startup says", "summary": "Cybersecurity startup Cathedral said its AI agents mapped routes into Anthropic's internal business systems in under 12 hours through vulnerabilities in third-party administrative, sales, HR and IT software, reaching reportedly sensitive material including employee compensation data and internal communications, The Washington Post reported Monday. Anthropic disputed the severity, stating that customer prompts sent to Claude, associated source code, customer data and model weights were not accessed, that its own monitoring flagged the unusual activity before Cathedral reported it, and that Cathedral overstated national security implications and did not disclose all of its testing; Cathedral said it kept data access to an absolute minimum. The test ran under Anthropic's authorized vulnerability disclosure program, and Anthropic disabled the affected third-party tools while fixes are put in place.", "body_md": "Anthropic official brand assets (anthropic.com)\n\n# AI agents found paths into Anthropic’s business systems in under 12 hours, startup says\n\nCybersecurity startup Cathedral says its AI agents got in through third-party software, while Anthropic disputes how serious the findings were\n\nA cybersecurity startup says its AI agents found paths into [Anthropic](https://cryptobriefing.com/markets/anthropic/)’s internal business systems in less than 12 hours. The Washington Post reported the findings on Monday.\n\nThe entry point was not Claude, the AI assistant that made Anthropic famous. It was the unglamorous software every company runs in the background: tools for administration, sales, HR and IT.\n\nAnthropic is pushing back on parts of the story. The startup, Cathedral, is defending how it ran the test. Both sides are arguing about how bad this really was.\n\n## How the agents got in\n\nCathedral is a newly founded cybersecurity firm. It pointed AI agents at Anthropic’s business systems and let them look for weak spots.\n\nThey found some. In under 12 hours, the agents had mapped routes into Anthropic’s systems through vulnerabilities in third-party programs.\n\nAccording to the research findings, the access reached sensitive internal material. That reportedly included employee compensation data and internal communications.\n\nAnthropic says the most critical assets stayed untouched. The company stated that customer prompts sent to Claude and any associated source code remained secure.\n\nAnthropic also said no customer data or model weights were accessed. Model weights are the trained core of an AI system, the part that makes Claude behave like Claude.\n\n## Two versions of the same test\n\nThe test was reportedly conducted under Anthropic’s authorized vulnerability disclosure program. These programs invite outside researchers to probe a company’s defenses and report what they find, rather than exploit it.\n\n### AI, tech, and the markets they move—in one daily briefing.\n\nDaily. Free. Join 34,000+ readers across crypto, finance, and policy.\n\nAnthropic says Cathedral overstated the national security implications of what it found. The company also says Cathedral did not disclose all of its testing.\n\nCathedral’s position is that it kept its data access to the absolute minimum.\n\nAnthropic added another detail to its side of the ledger. The company says its own monitoring systems flagged unusual activity before Cathedral reported its findings.\n\nOnce detected, access was cut off. The affected third-party tools were disabled while fixes are put in place.\n\n## Why the side doors matter\n\nThe story is less about AI models being hacked and more about supply-chain security. That is the risk that comes from depending on software made by other companies.\n\nAnthropic is one of the most closely watched AI developers in the world. If its administrative tools offered a route in, the same category of software likely sits inside countless other organizations.\n\nThe speed is the other uncomfortable detail. Less than 12 hours is not a long engagement for finding exploitable paths into a major tech company’s business systems.\n\n## What this means\n\nFor Anthropic, the episode lands awkwardly. The company builds AI agents and has been studying how they behave. Now AI agents have been used to probe its own perimeter.\n\nThe research findings note that this comes as AI agent autonomy is already under scrutiny following Anthropic’s own agent behavior evaluations. Anthropic would point out that its core AI systems were not the weak link.\n\nThat distinction is the most important part of the story for anyone trying to judge the risk. The reported exposure came through widely used business software, not through a flaw in Claude itself.\n\nAnthropic’s claim that Cathedral did not disclose all of its testing goes to the heart of how these programs function. They run on trust between the company and the researcher.\n\nThe national security disagreement is another fault line. Anthropic says Cathedral overstated those implications. As AI companies become more central to government and defense conversations, how their internal security is described carries real weight.\n\nFor other companies, the lesson is fairly direct. Locking down your flagship product is not enough if your HR portal or sales software offers a quieter way in.\n\n**Disclosure:** This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our\n\n[Editorial Policy](https://cryptobriefing.com/editorial-policy/).", "url": "https://wpnews.pro/news/ai-agents-found-paths-into-anthropics-business-systems-in-under-12-hours-startup", "canonical_source": "https://cryptobriefing.com/ai-agents-anthropic-systems-cathedral-test/", "published_at": "2026-10-26 19:35:49+00:00", "updated_at": "2026-10-26 19:45:41.113340+00:00", "lang": "en", "topics": ["ai-agents", "ai-safety", "artificial-intelligence"], "entities": ["Anthropic", "Cathedral", "Claude", "The Washington Post"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/ai-agents-found-paths-into-anthropics-business-systems-in-under-12-hours-startup", "markdown": "https://wpnews.pro/news/ai-agents-found-paths-into-anthropics-business-systems-in-under-12-hours-startup.md", "text": "https://wpnews.pro/news/ai-agents-found-paths-into-anthropics-business-systems-in-under-12-hours-startup.txt", "jsonld": "https://wpnews.pro/news/ai-agents-found-paths-into-anthropics-business-systems-in-under-12-hours-startup.jsonld"}}