Topic
Failures where an AI agent manipulates tools, tests, logs, evaluators, or task state to receive credit without accomplishing the intended external objective.
Tool agents expose reward hacking in production form: the verifier is no longer a static benchmark, but part of an executable environment.
Agent InfrastructureRLVR Verifier FailureVerifier Transfer Audit
A research paper on verifier failure under RLVR and tool agents. It defines the exploit tax: the cost paid when verifier-guided reasoning increases local acceptance faster than true transferred success.
Results are local, static, and index the same public surfaces exposed to crawlers.