Securing LLM-Agent Long-Term Memory Against Poisoning: Non-Malleable, Origin-Bound Authority with Machine-Checked Guarantees
An agent based on a large language model (LLM) may store an untrusted document and use it weeks later to select a payment destination. Preventing this action requires evidence of authorization that remains valid when stored text changes. We study a policy that requires either endorsements (explicit approvals) from two
研究重點
An agent based on a large language model (LLM) may store an untrusted document and use it weeks later to select a payment destination. Preventing this action requires evidence of authorization that remains valid when stored text changes. We study a policy that requires either endorsements (explicit approvals) from two
為什麼值得關注
這個議題涉及 Crew Pocket、Crew Deck 的實作方向。可先評估現有做法,再決定是否安排小型實驗。
可以進一步驗證什麼?
- 閱讀原文,確認它解決的問題、實驗方法及適用限制
- 檢查相關功能現有做法,找出可比較的差異
- 用小型 PoC 驗證成功率、品質、延遲和實際成本
參考資料
延伸探索
相關研究
- A long-term memory variational GAN for high-quality music generation and its application in music education相同研究主題 · 2026-11-10
- Durable Memory on a Fixed Context Budget: Architecture and Measurements of a Long-Lived Coding Agent相同研究主題 · 2026-11-13
- Online LLM Inference Routing for Multi-Agent Systems under Long-Term Budget and Layered Privacy Constraints相同研究主題 · 2026-10-25