Research Demonstrates Persistent Prompt Injection via LLM Long-Term Memory

# Research Demonstrates Persistent Prompt Injection via LLM Long-Term Memory

https://gcdradar.vercel.app/incidents/research-demonstrates-persistent-prompt-injection-via-llm-lo · AIR-2026-1465 · closed 2026-11-05

- Lane: Research
- Severity: 1 (Research / proof of concept)
- Fix status: Fix status unknown
- First reported: 2026-10-22
- AI role: Attack on AI
- Products: LLM agents

## Summary

Researchers demonstrate how persistent memory features in conversational AI enable cross-session prompt injection attacks and compromise. The work highlights new security risks when combining long-term memory with large language models.

## What to do

No action recorded; read the sources below and the vendor's advisory.

## Latest updates

- 2026-10-22: Combining Long-Term Memory and LLMs Comes at a High Cost: From Prompt Injection to Persistent Compromise (https://arxiv.org/abs/2610.28810)

## Sources

- 2026-10-22 arXiv: Combining Long-Term Memory and LLMs Comes at a High Cost: From Prompt Injection to Persistent Compromise https://arxiv.org/abs/2610.28810

Verify against the vendor's advisory before changing production. Data CC BY-SA 4.0, https://gcdradar.vercel.app.