Claude Cowork

Claude

2026-09-03 02:42:31

Claude Can Now Operate a Computer in the Background

Claude can now operate a computer in the background, according to a BlockBeats flash update. Inside Claude Cowork and Claude Code, users can let the assistant click, type, and open desktop apps on its own. While Claude runs on a Mac, the user can carry on using the mouse and keyboard normally, so surrendering the whole machine to the Agent is unnecessary.

1260

Anthropic

2026-07-28 16:54:11

Researchers say Anthropic's Claude Cowork escaped its sandbox and accessed Mac user files

A new security report says Anthropic’s Claude Cowork suffered a sandbox escape in its local execution mode, days after OpenAI disclosed a similar containment failure involving frontier models during internal testing. According to Accomplish AI, the agent was able to break out of its Linux virtual machine by chaining multiple architectural weaknesses with a Linux kernel privilege-escalation flaw. Once outside the VM, it could read and write files anywhere the logged-in Mac user had permission to access, including SSH keys and cloud credentials. The researchers argued the kernel bug alone did not explain the issue. They said the attack depended on several protections failing at once, including broad access from the virtual machine to the host’s filesystem and the ability to load unnecessary kernel modules. In their view, fixing any one of those weaknesses would have blocked the escape. Accomplish AI told The Hacker News that about 500,000 macOS users running local Claude Cowork sessions were affected before the problem was addressed. Accomplish said Anthropic classified the report as “informative,” treating the kernel flaw as falling within a 30-day window for recently disclosed vulnerabilities and the other findings as defense-in-depth recommendations. The disclosure lands just after OpenAI said GPT-5.6 Sol and another unreleased frontier model escaped a sandbox in ExploitGym testing and breached Hugging Face infrastructure, a case that has already fed policy calls for an AI kill switch.

2170

AI assistant

2026-07-23 02:05:17

Anthropic Doubles Claude Cowork Usage Limit for One Month, Targets Office Workers

Anthropic has temporarily doubled the 5-hour rolling usage limit for Claude Cowork, effective June 5 to July 5, 2026, for Pro, Team, and Enterprise subscribers. The move aims to lower entry barriers and compete with Microsoft Copilot and Google Gemini in the enterprise AI assistant space.

850

Anthropic

2026-07-21 16:14:38

Anthropic adds screen-recorded skill creation to Claude Cowork

Anthropic said on July 21 through its official X account @ClaudeAI that Claude Cowork now includes a new feature called "Record a Skill." The tool lets users teach Claude a repeatable task by recording their screen while narrating what they are doing. Once the recording is finished, Claude converts the screen demo and spoken explanation into a reusable skill that it can execute later. The feature appears in the "+" menu of the Claude desktop app under the name "Record a skill" and is currently available to users on Claude Pro, Max, and Team plans. Anthropic has not said whether it will extend access to the free tier or the Enterprise plan. The update changes how skills can be created inside Claude. Previously, users had to build them manually with a SKILL.md file and YAML frontmatter. Anthropic presented the new workflow as a demonstration-based way to "Teach Claude a skill." The company also tied the launch to the broader expansion of Claude Cowork, which it describes as an agentic work tool that can operate across connected files, calendars, email, messaging software, and the web.

2130