Unverified50% confidenceFactExact time
Memory poisoning is an evolved form of prompt injection attack, in which an attacker induces an AI to write false information into long-term memory storage, thereby influencing the system's subsequent interactions with all users.
1
Sources
50%
Confidence
Long-term
Relevance
10/5/2026
First Seen
Sources
Related Entities
Related Claims
Unverified攻击者常借助混淆技术隐藏恶意脚本诱导用户复制粘贴运行，这类手法被称为剪贴板劫持或社会工程学注入71% similarUnverifiedEntering passwords into an AI chat box is an extremely dangerous behavior, because user input is typically transmitted and stored as conversation history on the service provider's servers, posing a risk of exposure.69% similarUnverified当推理引擎存在内存安全漏洞时，模型生成的特定输出内容可能触发这些漏洞，使输出从数据变成可执行的攻击载荷69% similarUnverifiedA heap overflow is a memory safety vulnerability in which an attacker corrupts adjacent memory structures by writing data beyond the boundaries of allocated heap memory, potentially achieving arbitrary code execution or control of program flow.69% similarVerifiedPrompt injection attacks induce AI agents to execute malicious operations by embedding text disguised as system instructions in code comments, documentation, web content, or even file names.68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/976536API
curl https://kongchang.com/api/v1/knowledge/claims/976536MCP
get_claim(id=976536)