# Preprint demonstrates chat-based poisoning of reused AI reasoning caches

_Published Monday, October 19, 2026 at 2:15 AM EDT · Science, Security · Latest · Tier 2 — Notable_

Researchers report a chat-based attack that poisons stored reasoning reused by language models, achieving a 46% attack success rate in a reasoning-trace retrieval system. Their preprint describes R²ECAP, which requires only access to a system's chat interface.

The attack creates two reasoning traces. The first steers the model toward an adversarial target; the second retains that influence in cached internal data even after the first trace and its cache entries are discarded.

Tests covered three open-weight models and three datasets, with attack success reaching up to 79%. A popular prompt injection detector flagged only 1% of retained responses. The findings concern systems that reuse reasoning caches.

## Sources

- [cs.CR updates on arXiv.org](https://arxiv.org/abs/2610.22873)

---
Canonical: https://techandbusiness.org/newswire/lzue2VV6D1_JUfRZMNva5c
Published: 2026-10-19T06:15:57.943Z
Story chronology: 2026-10-19T04:00:00.000Z
Retrieved: 2026-10-19T08:38:53.999Z
Publisher: Tech & Business (techandbusiness.org)