OWASP

3 articles tagged with this topic

OWASP

AI Agents Get to Work—OWASP Tops 2025 LLM Risk List With Excessive Agency

OWASP's 2025 LLM Top 10 names Excessive Agency the #1 Agent risk—from 'saying wrong' to 'doing wrong.' Not taxonomy; engineering teams must answer.

Aug 192 min readAI Finance Weeklyjuejin.cn

Agent

AI agents now write code and change databases. OWASP shows 41% of LLM flaws come from over-permissioning. Sandboxing is governance, not engineering.

Aug 143 min readAI Finance Weeklyjuejin.cn

M CP

Research finds 492 public MCP servers vulnerable; CVE-2025-6514 affects 437,000+ downloads across production deployments.

Apr 184 min readopcnew Editorialjuejin.cn