Microsoft Excel security flaw enables data theft with Copilot

A security vulnerability in Microsoft Excel combines spreadsheets and the Copilot Agent to potentially steal data, according to TechRadar. The flaw is described as fascinating in its approach. The report was published on March 11, 2026.

TechRadar has highlighted a security issue in Microsoft Excel where spreadsheets team up with the Copilot Agent to facilitate data theft. The article's title refers to this vulnerability as 'fascinating,' suggesting an innovative or unexpected method of exploitation.

The description provided in the report states, 'There's more than one way to skin an Excel table,' implying creative uses of Excel's features that go beyond standard functionality. This points to how the software's capabilities might be misused in conjunction with AI tools like Copilot.

No further technical details or timelines are specified in the available information, but the publication date of the article is March 11, 2026, at 17:30 UTC. This discovery underscores ongoing concerns about data security in productivity software integrated with AI agents.

相关文章

Realistic photo of a Windows 11 laptop showcasing advanced Microsoft Copilot AI features, including voice activation and screen analysis, in an office environment.
AI 生成的图像

Microsoft brings advanced Copilot AI to all Windows 11 PCs

由 AI 报道 AI 生成的图像

Microsoft has announced a series of generative AI features for Windows 11, aiming to transform every PC into an 'AI PC' through voice activation, screen analysis, and automated file handling. These updates, including the 'Hey, Copilot' voice command and worldwide rollout of Copilot Vision, build on the company's agentic AI focus. The features emphasize natural user interactions while addressing past privacy concerns from tools like Recall.

Security firm Varonis has identified a new method for prompt injection attacks targeting Microsoft Copilot, allowing compromise of users with just one click. This vulnerability highlights ongoing risks in AI systems. Details emerged in a recent TechRadar report.

由 AI 报道

Microsoft has issued an emergency patch for a worrying security flaw in its Office software. The vulnerability could allow hackers to access users' files if not updated promptly. The patch was released to address this critical issue.

Researchers have identified three high-risk vulnerabilities in Claude.ai. These enable an end-to-end attack chain that exfiltrates sensitive information without the user's knowledge. A legitimate Google ad could trigger data exfiltration.

由 AI 报道

Security researchers, first reporting via TechRadar in December 2025, warn WhatsApp's 3 billion users of GhostPairing—a technique tricking victims into linking attackers' browsers to their accounts, enabling full access without breaching passwords or end-to-end encryption.

中国国家网络安全机构警告OpenClaw AI代理软件存在安全漏洞,可能允许攻击者完全控制用户计算机系统。该软件最近下载量激增,主要云平台提供一键部署服务,但默认安全配置薄弱。

由 AI 报道

A TechRadar report states that over 29 million secrets were leaked on GitHub in 2025. The article suggests that AI is not helping and may be making the situation worse.

 

 

 

此网站使用 cookie

我们使用 cookie 进行分析以改进我们的网站。阅读我们的 隐私政策 以获取更多信息。
拒绝