Evidence · DiggingBeagle record
AutoRedClaw preprint, arXiv 2610.27244
- Source role
- primary disclosure
Evidence record
Claim-level citations (1)
- supportsAutoRedClaw evaluates persistent-state attacks across agent harnesses: AutoRedClaw reports a 330-case benchmark across seven harnesses and four models, with most tested agent systems exceeding 50% attack success.
Abstract; persistent configuration, memory and skill surfaces; RedClaw-Bench evaluation