#memoryshapes
Whose Credential Is It? How Owner Attribution in Agent MemoryShapes Cross-User Account Misuse

Hau-Sing So et al.

#arXiv #cs.AI #cs.CL
Whose Credential Is It? How Owner Attribution in Agent MemoryShapes Cross-User Account Misuse
Shared long-term memory lets an LLM agent serving one user inadvertently retrieve another user's credentials, but it is unclear when such exposure leads to actual harm. We introduce a three-stage audit that separates three events: a foreign credential reaches the prompt (exposure), the agent acts w…
arxiv.org
October 21, 2026 at 12:04 AM
Whose Credential Is It? How Owner Attribution in Agent MemoryShapes Cross-User Account Misuse

Hau-Sing So et al.

#arXiv #cs.AI #cs.CL
Whose Credential Is It? How Owner Attribution in Agent MemoryShapes Cross-User Account Misuse
Shared long-term memory lets an LLM agent serving one user inadvertently retrieve another user's credentials, but it is unclear when such exposure leads to actual harm. We introduce a three-stage audit that separates three events: a foreign credential reaches the prompt (exposure), the agent acts w…
arxiv.org
October 20, 2026 at 10:34 PM
Hau-Sing So, Daixian Liu, Carl Hou, Yuhaohang He, Mengzhu Wang, Hai-Tao Zheng: Whose Credential Is It? How Owner Attribution in Agent MemoryShapes Cross-User Account Misuse https://arxiv.org/abs/2610.25707 https://arxiv.org/pdf/2610.25707 https://arxiv.org/html/2610.25707
October 20, 2026 at 6:40 AM