Unverified50% confidenceFactExact time
Direct prompt injection requires the attacker to interact with the model directly and is relatively easy to intercept with input-layer filtering.
1
Sources
50%
Confidence
Long-term
Relevance
10/1/2026
First Seen
Sources
Related Entities
Related Claims
VerifiedPrompt injection attacks can use carefully crafted user inputs to induce the model to ignore the constraints of the system prompt.75% similarUnverifiedAttackers 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.72% similarUnverifiedIndirect 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.69% similarUnverifiedThread replies allow Agents to respond precisely at the point where the conversation occurs, keeping channels clean and consolidating context.68% similarUnverified在多人协作线程中，提示注入（Prompt Injection）的攻击面扩大，任何有访问权限的用户都可能成为注入向量，恶意指令可能被伪装成背景补充或方向修正67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/966194API
curl https://kongchang.com/api/v1/knowledge/claims/966194MCP
get_claim(id=966194)