MutMem‑V2 builds on the original MutMem‑V1 by providing a complete portable verification contract and a clean‑install reproduction path for persistent agent memory. It defines exact canonical bytes, domain‑separated commitments, mandatory recall‑evidence, and other security primitives, and releases a protocol with 18 object schemas, 39 recall vectors, 15 mutation vectors, and 37 closed recall failure reasons. Independent Node and Python implementations agree on all 72 structural and cryptographic terminals, and a production‑conformance corpus confirms 42/42 cases across 28 required classes, demonstrating reproducible evidence and portable integrity under stated assumptions.
MutMem V1 introduced retention-preserving, cryptographically authorized mutation for persistent agent memory but did not provide a complete portable verification contract or clean-install reproduction...
arXiv:2608. 02843v1 Announce Type: cross Abstract: Persistent agent memory must adapt as later outcomes change earlier evidence, yet mutable retrieval weights create an attribution problem: reviewers must distinguish authorized adaptation from database tampering.
arXiv:2604. 16548v2 Announce Type: replace-cross Abstract: The emergence of writable, cross-session persistent memory in LLM agents introduces a qualitatively different threat landscape from conventional input-centric security concerns, characterized by three properties: persistence, statefulness, and propagation.
arXiv:2607. 23929v1 Announce Type: new Abstract: LLM agents increasingly coordinate through persistent shared memory: one agent's write becomes another agent's premise, and eventually a tool call with real side effects.
arXiv:2609. 00546v1 Announce Type: cross Abstract: Agent systems are commonly described by the model and harness that currently produce their behavior.
arXiv:2609. 04875v1 Announce Type: cross Abstract: Long-running LLM agents are stateful: beyond the transcript they accrete compressed summaries, plaintext memory, pending tool plans, and, under every serving API, a KV cache.
arXiv:2607. 27080v1 Announce Type: cross Abstract: Memory systems allow agents to retain and reuse information from past interactions, but they can also let malicious content persist.
The paper introduces the concept of Cognitive Serializability for autonomous AI agents, ensuring that mutations derived from dynamic inputs—such as database reads, evidence, policy, beliefs, and delegated authority—are committed in a serial, logically consistent order. It presents a framework called Trusted Cognitive Transaction (TCT) that combines immutable executable definitions, sealed envelopes, guard-first commits, and receipt-driven reconciliation to enforce serializability and prevent anomalies. Experimental results show that the prototype implementation incurs minimal overhead while eliminating injected anomalies.
arXiv:2606. 01138v2 Announce Type: replace-cross Abstract: Agent-memory frameworks -- mem0, Letta/MemGPT, Cognee, Zep/Graphiti, MemoryOS, MemTensor -- each ship their own SDK, storage layout, and operational vocabulary.
arXiv:2606. 15903v1 Announce Type: cross Abstract: Where an LLM sits in an agent memory pipeline -- between the recall plane that retrieves stored facts (extensively benchmarked) and the control plane that mutates them via supersede, release, purge (largely untested) -- shapes which forgetting failure modes the system recovers.
arXiv:2609.22243v1 Announce Type: new Abstract: Input-conditioned neural interventions raise a runtime question: what persists when one behavioral specification admits multiple actions whose validity...
arXiv:2608.21867v1 Announce Type: new Abstract: LLM agents are moving from single-prompt use to long task streams in which reusable memory becomes a core capability for terminal, software-engineering...