Unverified50% confidenceFactExact time
After an account is maliciously taken over, the attacker can read historical conversations and may initiate new interactions as the user to conduct social engineering attacks or abuse the service.
1
Sources
50%
Confidence
Long-term
Relevance
10/31/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedOpenCode explained that the attackers do not directly register new accounts; instead they first steal a batch of "clean" accounts with an existing history of trust, then use these accounts to launch distillation requests, so an attack can persist for a period before being detected and banned.72% similarVerifiedAttackers can induce agents to execute a sequence of seemingly normal operations—such as querying user data, calling formatting tools, and triggering email sending—to form a data exfiltration chain.72% similarUnverifiedWhen a maintainer's account is compromised, attackers may plant malicious code in the project and distribute it to all downstream users through subsequent updates.71% similarUnverifiedIn large language model services, the consequences of ATO take on new particularities: attackers can read historical conversations (which may contain code, documents, trade secrets, personal privacy), can call the model API as the victim to consume quota or trigger charges, and can even exploit the model's "memory" feature to poison the context.70% similarUnverified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69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1119941API
curl https://kongchang.com/api/v1/knowledge/claims/1119941MCP
get_claim(id=1119941)