Unverified50% confidenceFactExact time
Prompt injection can be divided into two types: direct injection and indirect injection, where indirect injection refers to attackers hiding malicious instructions in external content read by the agent.
1
Sources
50%
Confidence
Long-term
Relevance
9/21/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedPrompt injection induces agents to deviate from the original task intent by embedding text disguised as instructions within external content processed by the agent.78% similarUnverifiedThe essence of prompt injection is that attackers hijack the execution logic of large language models through carefully crafted text content.74% similarUnverifiedAttackers may use geometric alignment methods to map anonymized embedding vectors back to an interpretable semantic space, potentially partially recovering the original text content.71% similarUnverified提示词注入分为直接注入和间接注入两种形式，因语言模型无法区分数据与指令，凡进入上下文窗口的文本都可能被执行69% similarUnverifiedIn concurrently scaled multi-agent systems, read-write race conditions occur where one agent reads shared state before another has finished writing.68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/938867API
curl https://kongchang.com/api/v1/knowledge/claims/938867MCP
get_claim(id=938867)