Unverified50% confidenceOpinionExact time
There is no reliable 'detect and block' solution for prompt injection, because it is difficult to judge whether a piece of text is an ordinary description or a malicious instruction, and to the model the two are identical in form; isolation and permission control are more realistic than content filtering.
1
Sources
50%
Confidence
Long-term
Relevance
10/15/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedFilter/detection-based defenses (such as scanning for phrases like 'ignore previous instructions') are structurally doomed to fail, because natural language has no syntax that distinguishes instructions from data.79% similarUnverifiedPrompt injection exploits not a vulnerability in the code parser, but the model's own insufficient awareness of the boundary between instructions and data74% similarUnverified用户可以通过提示注入和语义改写等手段绕过内容检测，使实际拦截效果是概率性的而非绝对的74% similarUnverifiedJudging how semantically close two pieces of text are cannot be done well with hand-written rule-based programs; it requires a large model to perform comparative scoring.73% similarUnverifiedTraditional code review processes often struggle to distinguish AI-generated comments.73% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1036879API
curl https://kongchang.com/api/v1/knowledge/claims/1036879MCP
get_claim(id=1036879)