Unverified50% confidenceFactExact time
Agents dynamically decide their next action based on context, and their behavior paths are generated only at runtime, breaking the traditional security model's assumption that the behavior of an executing entity can be enumerated.
1
Sources
50%
Confidence
Long-term
Relevance
10/21/2026
First Seen
Sources
Related Entities
Related Claims
VerifiedThe agent loop follows a 'reasoning-action-observation' paradigm, iterating repeatedly until the goal is achieved or a stopping condition is reached.78% similarUnverifiedPolicy drift refers to an agent gradually deviating from its originally configured behavioral boundaries over the course of multi-turn conversations or complex reasoning chains.76% similarUnverifiedThere is a time window between when the policy check passes and when the executor carries out the action, during which state drift, action tampering, and replay attacks may occur.75% similarUnverifiedAn autonomous agent can execute a second action before a human or system has finished evaluating the first action, and this time window is where the vast majority of reported AI incidents originate.75% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1079887API
curl https://kongchang.com/api/v1/knowledge/claims/1079887MCP
get_claim(id=1079887)