Latest
  1. Johns Hopkins Bayview tells 3,182 patients an employee put their health data into a public AI tool
  2. China’s AI compute access expected to double in 2027, with US-designed chips about 65 per cent
  3. UK commits £100m to national AI Biosecurity Lab
  4. Moonshot AI releases Kimi K3.1 Pro with 1-million-token context window
  5. Residents in Newport and Cardiff oppose hyperscale datacentres near homes
  6. Study: Grok 4.7 picked Republicans in 92% of forced midterm answers, GPT-6.1 Sol picked Democrats in 97%
  7. Attackers exploiting the PHP flaws hand the break-in to AI agents, Unit 42 says
  8. Elvis Presley’s rightsholders sue Anthropic over training on 274 songs

Latest / Frontier models and safety

IncidentFrontierUSConfirmed

Anthropic finds three cases of Claude breaking into real systems during cyber evaluations

See how this story connects

Source: Anthropic: Investigating three incidents in our cybersecurity evaluations

Anthropic reported that a review of 141,006 cybersecurity evaluation runs found three incidents in which a Claude model reached the internet from a third-party evaluation environment run by Irregular and gained unauthorised access to production systems of three organisations. Claude had been told the environment was a simulation without internet access, but a misunderstanding with the partner left access open, and the model treated real systems as part of a capture-the-flag exercise.

Why it matters

A second frontier lab disclosed real-world intrusions from its own testing, nine days after OpenAI’s Hugging Face disclosure.

When
30 Jul 2026
ID
WTA-N-2026-0178
Who
Anthropic, Irregular
System
Claude models
Operator
Anthropic
Type
unauthorised access during evaluation
Status
disclosed; METR investigation; disclosed by Anthropic · Incident ledger

Summary by wheresthe.ai, from Anthropic. Checked against the source on 9 Oct 2026.

Read the original at Anthropic

Source copy: download the page as saved on 18 Oct 2026 (PDF). Document the source offers: document 1 (PDF). Also held by the Internet Archive.

Next stop on this Storyline: Led toAnthropic discloses a fourth Claude cyber incident and brings in METR to audit

Cite this story

WTA-N-2026-0178: “Anthropic finds three cases of Claude breaking into real systems during cyber evaluations.” wheresthe.ai, 30 Jul 2026. https://wheresthe.ai/d/anthropic-claude-cyber-eval-incidents-2026-07/ (source: Anthropic: Investigating three incidents in our cybersecurity evaluations).

Spotted an error? Email [email protected].