Unverified50% confidenceFactExact time
RAG applications concatenate retrieved text with instructions and feed them to the model together, causing attacker text and operator instructions to travel in the same channel.
1
Sources
50%
Confidence
Long-term
Relevance
10/19/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedAttackers tamper with a model's chat template by calling the create API, permanently injecting malicious instructions into the System prompt; this change is stored in the model's Modelfile and is invisible to the client and undetectable by standard tools.70% similarUnverifiedRAG技术使Agent能够突破单次对话的记忆限制，通过语义相似度检索召回相关内容实现跨会话的知识积累与调用70% similarUnverified程序通过余弦相似度计算相关段落，动态拼装发送给 AI 的 prompt，实现标准 RAG 流程70% similarUnverifiedRAG系统中承担检索职责的Embedding模型可以进行专门的微调70% similarUnverifiedDify has a built-in high-quality RAG engine; after users upload documents, the platform automatically completes text chunking, vector storage, and semantic retrieval.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1069264API
curl https://kongchang.com/api/v1/knowledge/claims/1069264MCP
get_claim(id=1069264)