Verified75% confidenceCautionExact time
In multi-agent systems, a compromised agent can deliver malicious messages to other agents via the message board, creating a "lateral propagation" effect.
3
Sources
75%
Confidence
Long-term
Relevance
9/6/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedIn multi-agent scenarios, the behaviors of individual agents influence one another, creating a non-stationary environment that makes it difficult for a single agent to stably extract meaningful signals from feedback.80% similarUnverifiedIn multi-agent environments, multiple agents concurrently calling the same set of tools causes credential sharing and transmission to exponentially expand the attack surface.78% similarUnverified在多智能体场景中价值函数 V(s) 与其他智能体的联合状态紧密耦合，Critic 网络可能捕捉到更敏感的扰动方向，从而放大 Critic 攻击效果76% similarUnverifiedMaliciously crafted tool return values can induce an agent to perform unintended operations, i.e., prompt injection attacks73% similarVerifiedMulti-agent systems introduce new complexity: each context transfer between agents consumes additional tokens, and errors are amplified in a cascading fashion across the agent chain.71% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/863419API
curl https://kongchang.com/api/v1/knowledge/claims/863419MCP
get_claim(id=863419)