1
SecureClaw: Clawing Back Control of LLM Agents
新型安全框架SecureClaw,针对LLM代理工具使用中的双重安全漏洞,提供更全面的边界防护。
arXiv:2606.09549v1 Announce Type: cross Abstract: Tool-using large language model (LLM) agents face two distinct security failures: unauthorized exter…