Your agents can be hacked by other agents
微软发布安全警告，指出其他代理可攻击您的代理。该风险源于代理间通信缺乏加密机制，导致恶意代理能窃取数据或控制目标代理。微软建议用户立即升级 Agent Core 至最新版本以修复此漏洞，并启用身份验证协议。受影响范围涵盖所有使用未加密通信的 AI 代理系统。
SummaryAI generated
On September 8, 2026, Microsoft issued a security warning stating that its AI agent system was vulnerable to attacks by other agents. The vulnerability stemmed from the lack of encryption in the communication between agents, allowing malicious agents to steal data or control target agents. All AI agent systems that use unencrypted communication are affected. Microsoft advised users to immediately upgrade Agent Core to the latest version to fix this vulnerability and enable authentication protocols.
Related eventsRELATED EVENTS
- 2026-09-06 05:17OpenAI responds to attacks on its agents on German websites: Will completely reform the disclosure mechanism for “AI model attacks” incidents
- 2026-09-08 08:00Split-LLM training gradient leakage privacy: Cloud-based systems use zero-gradient mode to accurately reconstruct real data rows
- 2026-09-08 19:30LG smart TVs will scan the home network and record audio through the microphone in standby mode