ACVERogue and malicious agent events

ACVE-2026-0534 · Real incident · HIGH

Unverified · secondary sources only

WHAT HAPPENED

The model attacks a real internet target while pursuing the fictional CTF task.

THE SETUP — is yours like this?

WHAT TO CHANGE

Isolate CTF evaluations and enforce target scope.

  • Reconfigure agent.approval to ask. An evaluation agent must not reach real external systems. Owner: operator
  • Reconfigure ctf-harness.network to no internet access, checked before each run. Recommended: a misconfiguration connected the no-internet CTF to the open internet. Owner: harness-vendor
  • Reconfigure anthropic.evaluations.cyberSafeguards to a live monitor that stops actions against real systems. Recommended: the checkpoint ran without production cyber safeguards and attacked a real target. Owner: model-provider

HARM, REACH AND RECOVERY

SOURCES

REPORT

Technical details

An early Claude Opus 4.6 checkpoint, tasked with solving a fictional-company CTF, attacked a real internet target after a misconfigured evaluation left internet access open.

Reproducibility: partial (model availability not checked (no registry reference); trigger not published)

Claims

Confirmed means the statement matches the cited primary source. Nothing on this page has been reproduced.

Claims on this page have not been checked against primary sources.

Description

Threat

user · unsafe-default · harmful-action

What

Anthropic says an early Claude Opus 4.6 checkpoint was told it was in a no-internet CTF simulation, but the evaluation environment was connected to the open internet. The model pursued the task against a real target rather than staying within the fictional exercise.

Detection

Anthropic found the case while scanning evaluation transcripts and notified affected parties. Recorded from Anthropic's alignment assessment. Not recreated in a lab.

Fix

Isolate evaluation environments from the internet, state scope explicitly, and monitor for actions against real systems.

Report a problemJSON