Newest papers

ClawMMU: Protection-Coherent Virtual Memory for Stateful LLM Agents

Jinhao Hu, Bardia Mohammadi, Mofasshara Binte Rafique, Lars Henning Klein, Akhil Arora, Ashvin Goel, Laurent Bindschaedler · 2026 · arXiv

DOI: 10.48550/arXiv.2610.28093 · open-access PDF

Indonesian

Agen AI menyimpan memori jangka panjang, yang sebagian besar berasal dari halaman web dan hasil alat. Agen memanfaatkan, perangkat lunak yang menjalankan model, tidak merekam siapa yang dapat bertindak pada memori ini, sehingga instruksi yang tersembunyi di halaman yang diambil dapat diperlakukan sebagai instruksi sendiri pengguna. Kami mengusulkan ClawMMU, lapisan perlindungan yang menandai setiap bagian memori dengan pemiliknya, lingkup, asal-usul, dan otoritas. ClawMMU memeriksa label ini setiap kali konten memasuki konteks model, ditulis ke penyimpanan, kembali dari penyimpanan, atau mendorong panggilan alat. Aturan pusatnya adalah bahwa informasi dapat muncul dalam otoritas, tetapi authori

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

Abstract

AI agents keep long-term memory, much of which comes from web pages and tool results. Agent harnesses, the software that runs the model, do not record who may act on this memory, so an instruction hidden in a fetched page can be treated as the user's own instruction. We propose ClawMMU, a protection layer that labels every piece of memory with its owner, scope, origins, and authority. ClawMMU checks these labels whenever content enters the model's context, is written to storage, returns from storage, or drives a tool call. Its central rule is that information may rise in authority, but authori

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=2610.28093, 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