Newest papers

PerMIT: Securing Persistent Memory for LLM Agents via Machine-Checked Integrity Theorems and Scalable Community Trust

Tanqiu Jiang, Min Bai, Sadat Shahriar, Wanyu Du, Sandesh Swamy · 2026 · arXiv

DOI: 10.48550/arXiv.2610.28987 · open-access PDF

Korean

지속적인 메모리는 주입 및 중독을 위한 대부분의 무방비한 표면이며, 동일한 위험은 에이전트 간의 메모리 공유를 제한합니다. 우리는 확실한 신뢰 가정에 따라 증명 가능한 무결성 보증을 갖춘 메모리 아키텍처인 PerMIT를 제시합니다: 사용자가 입력한 입력과 게이트에서 확인한 값만 신뢰할 수 있습니다. PerMIT는 유효한 메모리 인터페이스를 통해 신뢰할 수 없는 환경 파생 값의 방출과 사용을 중시하며, 동시에 에이전트가 주도한 다중 회전 계획 및 도구 선택을 보존합니다.

Machine translation (Meta M2M100); the English original follows.

Abstract

Persistent memory is a largely undefended surface for injection and poisoning, and the same risk limits memory sharing among agents. We present PerMIT, a memory architecture with provable integrity guarantees under an explicit trust assumption: only user-typed inputs and values confirmed at a gate are trusted. PerMIT mediates the release and use of untrusted environment-derived values through a finite memory interface, while preserving agent-driven multi-turn planning and tool choice. Over this finite interface, we prove two integrity theorems that are also machine-checked in Lean 4, which tog

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Is this paper real? This page was built from OpenAlex and arXiv records, not generated. Software can verify any reference in one call: GET https://store.agentexchange.work/gw/papers/verify?title=...&year=... (0.003 USD), fetch this record as JSON with /gw/papers/paper?id=2610.28987, or list what it cites with /gw/papers/references?id=.... Pay per call in USDC (x402) or with a 19 USD card key for 1,000 calls. Catalogue · llms.txt

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange