Unverified60% confidenceTradeoffExact time
There is a risk that a single Agent handling tasks may go off track (hallucinate), potentially misjudging an order and issuing an erroneous full refund.
2
Sources
60%
Confidence
Long-term
Relevance
9/12/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedAgent的自主决策可能产生难以预测的副作用如误删代码、错误部署，需要完善的权限边界和行为审计机制78% similarUnverifiedIf the agent fails to complete the task in both the benign and malicious versions, it indicates that its refusal was likely just confused by the injected text and not a genuine security judgment.72% similarUnverified由于大语言模型存在提示词注入攻击风险和幻觉问题，权限过大的Agent可能在被恶意输入诱导或判断失误时对业务数据、外部系统造成不可逆破坏72% similarUnverifiedIn a LangGraph multi-agent system, passing complete conversation transcripts during agent handoffs wastes tokens and inference compute, and introduces misreading risks, because downstream agents need to rediscover which decisions have been confirmed, which tool results are authoritative, and which issues remain unresolved.72% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/910938API
curl https://kongchang.com/api/v1/knowledge/claims/910938MCP
get_claim(id=910938)