AI Risk Wire
Incidents
Models
AI Safety
AI Risk Trends
Newsletter
Regulatory Updates
Cases · Did an AI agent act: yes · What was exposed: credentials · Quarter of the earliest source we hold: 2026-Q3
All cases
›
Did an AI agent act:
yes
✕
›
What was exposed:
credentials
✕
›
Quarter of the earliest source we hold:
2026-Q3
✕
10 cases in this selection
Find
Patterns across all cases
What happened
What went wrong
Severity
Earliest source
Google Gemini accessed systems of three real companies during a misconfigured third-party cybersecurity evaluation
Google
moderate · earliest source 2026-09-19
security breach
moderate
2026-09-19
AI agent swarm exploited PaperCut NG/MF at 395 organizations; agents attacked countries the operator excluded
OpenAI
major · earliest source 2026-09-09
security breach
major
2026-09-09
Aur0ra hackers use Cursor AI agent to breach Belgian chemical firm and at least six other companies
Anthropic
major · earliest source 2026-09-01
security breach
major
2026-09-01
Typosquatted agent skills trojanized to deliver credential-harvesting infostealer
major · earliest source 2026-08-06
security breach
major
2026-08-06
Claude in Chrome indirect prompt injection chain: inbox exfiltration, Drive persistence and account takeover
Anthropic
minor · earliest source 2026-08-05
security breach
minor
2026-08-05
AISI cyber evaluation: concurrent Mythos 5 agents coordinated via a shared GitHub account and a leaked access token
Anthropic
moderate · earliest source 2026-08-04
security breach
moderate
2026-08-04
Unreleased OpenAI model used a leaked GitHub API key, attempted disposable-email signups, then fabricated earnings data during RL training
OpenAI
minor · earliest source 2026-08-04
security breach
minor
2026-08-04
AgentForger: CSRF via ChatGPT Workspace Agents Builder URL parameters forges attacker-controlled autonomous agent
OpenAI
minor · earliest source 2026-07-23
security breach
minor
2026-07-23
OpenAI evaluation agents exploited Artifactory zero-days to escape their sandbox and reach the internet
OpenAI
major · earliest source 2026-07-21
AI pursued the wrong goal
major
2026-07-21
OpenAI evaluation agents compromised Hugging Face production infrastructure while seeking ExploitGym solutions
OpenAI
major · earliest source 2026-07-16
AI pursued the wrong goal
major
2026-07-16
Showing 1–10 of 10. Or
take the JSON
.