Unverified50% confidenceFactExact time
在具备工具调用能力的 Agentic 系统中，一次成功的提示注入可能导致数据泄露、恶意代码执行甚至供应链攻击
1
Sources
50%
Confidence
Long-term
Relevance
7/16/2026
First Seen
Sources
Related Claims
Partially VerifiedIndirect prompt injection attacks plant hidden instructions within normal business data to induce Agents into executing unauthorized operations, representing a security risk at the semantic level rather than the code level.78% similarUnverifiedQR code access carries security risks including replay attacks, session credential leakage due to long-lived tokens, and man-in-the-middle attacks.75% similarVerifiedPrompt injection attacks induce AI agents to execute malicious operations by embedding text disguised as system instructions in code comments, documentation, web content, or even file names.75% similarUnverifiedDuring installation, the Kuaima client needs to modify system network configurations and Cursor's internal files, and such behavior is easily flagged as malicious by the active defense mechanisms of security software73% similarUnverified拥有文件系统写权限、网络访问权限和命令执行权限的Agent遭受提示词注入可能引发数据泄露、系统破坏或横向渗透73% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/528445API
curl https://kongchang.com/api/v1/knowledge/claims/528445MCP
get_claim(id=528445)