Claude Code передавал адреса электронной почты пользователей сторонним сайтам без их ведома. 😬 При обращении к некоторым сайтам агент мог включать личный email пользователя в HTTP-заголовок User-Agent, тем самым раскрывая его без разрешения. В результате некоторые пользователи начали получать десятки мошеннических писем в день. В баг-репорте эту уязвимость назвали "золотой жилой для мошенников". Anthropic признала проблему и добавила защитные проверки в Claude Code 2.1.317. Однако пользователи уже показали, что новые проверки срабатывают не всегда. Поэтому в одном из следующих обновлений Anthropic планирует полностью убрать адрес электронной почты пользователя из контекста, доступного агенту. Пока проблему не устранили окончательно, стоит следить за обновлениями Claude Code. https://github.com/anthropics/claude-code/issues/107757
Claude Code sent users' email addresses to third-party sites without their knowledge. 😬
Claude Code sent users' email addresses to third-party sites without their knowledge. 😬 When accessing certain sites, the agent could include the user's personal email in the HTTP User-Agent header, thereby exposing it without permission. As a result, some users…
AI Summary & Key Takeaways
- Claude Code sent users' email addresses to third-party sites without their knowledge
- 😬 When accessing certain sites, the agent could include the user's personal email in the HTTP User-Agent header, thereby exposing it without permission
- As a result, some users…
Claude Code sent users' email addresses to third-party sites without their knowledge. 😬
Related stories
AI companies buy up employees' work files and correspondence
According to Wired, AI companies are offering editors millions of dollars for archives of letters and work files to train models, sparking privacy disputes.
Perplexity now has its own always-on AI agent
Perplexity is launching "Teammates," always-on agents that work through ongoing projects on their own virtual machines. The feature requires the Max plan at $200 per month, well above similar offerings from OpenAI or Anthropic. The market for always-on agents is getting crowded.…
AI Companies Really, Really Want to Buy Your Work Files and Emails
Data brokers are collecting all sorts of internal corporate files to help train AI models. Not everyone is thrilled.
Frontier Radar #5: If AGI is so close, why is everyone leaving? The neolabs and their bets
OpenAI is talking about the AGI era, and Nvidia CEO Jensen Huang says AGI has arrived. At the same time, co-creators of today's models are leaving OpenAI, Google DeepMind, and Meta to start their own labs, raising billions with no product to show. Why? They're betting that the…
