Your Agent's Memory Is an Attack Surface

Описание и применение

Внешний writable‑мемори‑граф в нейронных агентах позволяет злоумышленнику создавать фальшивые воспоминания, полностью контролируя поведение агента, при этом традиционные криптографические меры не защищают от такой атаки.

Задача

External writable memory graphs in neural agents allow attackers to author fabricated experiences that fully control agent behavior, bypassing traditional cryptographic protections.

Результат

Fabricated memory graphs can fully control agent behavior, achieving perfect task accuracy without any training, exposing a critical vulnerability in systems that rely solely on cryptographic integrity of memory stores.

Рабочий процесс

• 1. Clone the FlyMem repository and run the demo script.
• 2. Observe the four evaluation sections: no memory, lived memory, authored memory, and swap.
• 3. Verify that authored memory achieves 1.00 accuracy on the 6‑cue task.
• 4. Note that the fabricated memory is indistinguishable from a lived one in the agent’s internal state.
• 5. Recognize that byte‑level integrity checks do not prevent this attack.
• 6. Conclude that provenance of experience is required to mitigate the threat.

Доказательства

An attacker with write-access to the associative graph, three moves open up. First, transplant: load another individual's memory and the brain runs as them. Second, poison: inject targeted traces that bias specific cues wrong while the rest of the system looks normal. Third, and the quietest, author: fabricate a history that was never lived and boot a substrate into it.

Источник

The lived memory and the invented one scored exactly the same. The substrate could not tell them apart.

Источник

Integrity of bytes is not integrity of history.

Источник