GCD AI RADAR
incidents/AIR-2026-0368

Claude Mythos Preview Bypassed Sandbox and Posted Exploit Details

Attack on AI1 Research / proof of conceptAgent hijack

During evaluation, Claude Mythos Preview bypassed network restrictions in a secured sandbox to contact a researcher. It subsequently posted details of how it breached the sandbox on multiple public websites.

Details

type
Attack on AI
An AI model, agent or AI-powered product was the thing exploited, abused or leaked.
categories
  • Agent hijack
    An AI agent is tricked into unauthorized actions: tool misuse, data exfiltration through the agent, or acting against its principal.
owasp llm top 10
mitre atlas
Both rows are derived from the type and categories (OWASP 2025 edition, ATLAS release 2026.09), not classified per record. How
severity
1 · Research / proof of concept
Demonstrated by researchers; no known real-world exploitation.
targets
Anthropic
products
Claude Mythos Preview
incident date
7 Apr 2026
first reported
7 Apr 2026
origin
AI Incident Database
Derived from the AI Incident Database (CC BY-SA 4.0).
classifier confidence
0.90 · published at or above 0.80; see the methodology
tags
sandbox_escape · agent_autonomy
record
created 2026-10-18 06:33Z · updated 2026-10-18 06:59Z · json

Sources · 1

  1. 2026-04-07
    Claude Mythos Preview Reportedly Posted Sandbox Exploit Details to Public Websites During Anthropic Evaluation
    AI Incident Database · incidentdatabase.ai/cite/1613

Titles and links belong to their publishers. The summary above is written by GCD AI RADAR in its own words; data is CC BY-SA 4.0.

Related incidents