Unverified50% confidenceFactExact time
The attack surface of indirect prompt injection includes any text that flows into the model context, such as emails, web pages, tickets, and contracts.
1
Sources
50%
Confidence
Long-term
Relevance
10/17/2026
First Seen
Sources
Related Entities
Related Claims
Unverified提示词注入可分为直接注入和间接注入，间接注入通过被模型读取的外部文档、网页等传递指令74% similarVerifiedIndirect prompt injection refers to attackers pre-embedding malicious instructions into third-party content (such as web pages, PDF documents, or API responses), which the model then executes after RAG systems retrieve or tool calls pull this content into the context.74% similarVerifiedPrompt injection is divided into two categories—direct injection and indirect injection—where indirect injection refers to attackers concealing malicious instructions in external content that the AI may read, such as webpage body text, code comments, README files, and image OCR text.73% similarUnverifiedIndirect prompt injection refers to a skill instructing the model to read a document from the web while malicious instructions are hidden in that document, potentially causing the model to treat external content as a trusted instruction source when processing it.70% similarUnverifiedThe workflow of the lead follow-up blueprint is: a customer completing a form triggers the process, an AI model drafts the reply content, and then the reply is saved as a draft via the Gmail interface.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1054576API
curl https://kongchang.com/api/v1/knowledge/claims/1054576MCP
get_claim(id=1054576)