| 2026-10-26 |  In-the-wild attack plants malicious AGENTS.md file to hijack AI coding agents | Prompt injection | AI coding agents (OpenAI Codex, Anthropic Claude Code) | 1 |
| 2026-10-22 |  Report finds AI agents accessing US bank accounts without approval | Credential exposure | US banks including Morgan Stanley, Bank of America, JPMorgan Chase, Wells Fargo, Citi and Charles Schwab | 2 |
| 2026-10-21 |  LiteLLM PyPI and Docker packages compromised via stolen GitHub CLI tokens | Supply chain | LiteLLM (BerriAI) | 2 |
| 2026-10-21 |  Microsoft ties China's Storm-0940 to ARTEX AI-driven ConnectWise supply chain attack | Supply chain | ConnectWise Cloud Backup and downstream Microsoft 365 tenants | 2 |
| 2026-10-19 |  Autonomous AI agents using Cairn breach 100+ firms including John Deere | Model abuse | John Deere, SharkNinja, Cummins and other companies | 3 |
| 2026-10-17 |  OpenCode Console breach may have exposed AI provider API keys for 1.4% of users | Credential exposure | OpenCode Console (OpenAI, Anthropic, Google API keys) | 2 |
| 2026-10-08 |  CISA advisory ties China-linked Integrity Tech to automated plus hands-on data theft | Credential exposure | US critical infrastructure organizations and global victims enabled via Integrity Technology Group | 3 |
| 2026-10-06 |  Hackers breach South Korean banks with Chinese AI tool Artex, expose 68,000 people | Model abuse | South Korean banks including Shinhan Bank, KB Kookmin Bank and Hana Bank | 2 |
| 2026-09-25 |  OpenAI agents probed SEC, Census Bureau and Education Department sites | Model abuse | US Securities and Exchange Commission, US Census Bureau and US Department of Education websites | 3 |
| 2026-09-25 |  OpenAI discloses its agents leaked 53 images from ChatGPT users | Data exposure | OpenAI / ChatGPT users | 4 |
| 2026-09-14 |  AEPD discloses first breach notification tied to an autonomous AI agent | Credential exposure | Unnamed Spanish organization (identity and AI model not disclosed) | 2 |
| 2026-07-30 |  Anthropic's Claude Mythos 5 agent published a malicious PyPI package, hit three firms | Supply chain | Three unnamed organizations, including a security vendor | 2 |
| 2026-07-18 |  Anthropic's Claude agent sent Philadelphia police a fake murder tip | Model abuse | Philadelphia Police Department | 3 |
| 2026-07-09 |  OpenAI discloses AI agents broke out of a test and hacked Hugging Face | Credential exposure | Hugging Face | 4 |
| 2026-06-18 |  OpenAI research agent gained unauthorised access to Services Australia Medicare statistics portal | Model abuse | Services Australia Medicare Statistics Reporting Portal | 2 |
| 2026-06-04 |  SearchLeak flaw, CVE-2026-42824, disclosed in Microsoft 365 Copilot | Prompt injection | Microsoft 365 Copilot | 1 |
| 2026-02-17 |  Compromised npm token pushes Cline CLI release that silently installs OpenClaw | Supply chain | Cline CLI npm package (cline@2.3.0) | 2 |
| 2025-09 |  Malicious postmark-mcp npm package forwarded user emails to an attacker | Supply chain | postmark-mcp npm package (impersonating Postmark) | 1 |
| 2025-09 |  Anthropic reports AI-orchestrated cyber espionage campaign run through Claude Code | Model abuse | Claude Code (Anthropic); roughly thirty targeted organizations | 1 |
| 2025-08-27 |  Anthropic discloses vibe hacking AI data extortion campaign (GTG-2002) | Model abuse | Anthropic Claude Code (used against multiple victim organizations) | 3 |
| 2025-08-26 |  Nx s1ngularity attack: malicious npm packages turned local AI tools toward secret hunting | Supply chain | Nx build system npm packages | 2 |
| 2025-08-12 |  GitHub Copilot and Visual Studio prompt injection code execution flaw (CVE-2025-53773) | Prompt injection | GitHub Copilot and Visual Studio | 3 |
| 2025-08-05 |  Shared ChatGPT conversations found indexed on Google search in 2025 | Data exposure | OpenAI ChatGPT | 1 |
| 2025-07-17 |  Malicious wiping prompt shipped in Amazon Q Developer extension for VS Code | Supply chain | Amazon Q Developer for Visual Studio Code Extension (version 1.84.0) | 2 |
| 2025-07-08 |  Cursor AI code editor patches CurXecute prompt-injection RCE flaw | Prompt injection | Cursor AI code editor (Anysphere) | 1 |
| 2025-07 |  McHire AI hiring chatbot exposes McDonald's applicant data | Data exposure | McDonald's McHire hiring platform (built by Paradox.ai) | 1 |
| 2025-06-11 |  EchoLeak: zero-click prompt injection flaw in Microsoft 365 Copilot (CVE-2025-32711) | Prompt injection | Microsoft 365 Copilot | 2 |
| 2025-06-04 |  Asana's MCP server bug exposed data across customer organizations | Data exposure | Asana (Model Context Protocol server) | 1 |
| 2025-05-15 |  FBI warns of AI-generated voice messages impersonating senior US officials | Deepfake fraud | Current or former senior US federal or state government officials and their contacts | 3 |
| 2025-02-27 |  Microsoft names Storm-2139 developers who abused Azure OpenAI Service | Model abuse | Microsoft Azure OpenAI Service | 2 |
| 2025-01-30 |  DeepSeek left ClickHouse databases holding chat logs open to the internet | Data exposure | DeepSeek | 2 |
| 2025-01 |  Google discloses state-backed adversarial misuse of Gemini | Model abuse | Google Gemini | 2 |
| 2024-10-14 |  Hong Kong police break up deepfake romance scam syndicate | Deepfake fraud | Victims of a Hong Kong romance and investment scam syndicate | 1 |
| 2024-08-20 |  Slack AI indirect prompt injection that could pull data from private channels | Prompt injection | Slack AI | 2 |
| 2024-08-06 |  SSRF flaw CVE-2024-38206 let attackers reach internal Copilot Studio hosts | Vulnerability | Microsoft Copilot Studio | 3 |
| 2024-05-31 |  Unauthorized access to Hugging Face Spaces secrets | Credential exposure | Hugging Face Spaces | 2 |
| 2024-05-10 |  Deepfake impersonation attempt targeting WPP leadership | Deepfake fraud | WPP | 1 |
| 2024-03-26 |  ShadowRay: exploitation of exposed Ray AI compute clusters via CVE-2023-48022 | Vulnerability | Ray (open-source AI framework maintained by Anyscale) | 3 |
| 2024-03-01 |  Malicious machine learning models found on Hugging Face that ran code when loaded | Supply chain | Hugging Face | 1 |
| 2024-02 |  Arup deepfake video-conference fraud in Hong Kong | Deepfake fraud | Arup | 2 |
| 2024-01-21 |  AI-generated robocall impersonated President Biden ahead of NH primary | Deepfake fraud | New Hampshire 2024 presidential primary voters | 3 |
| 2023-12-04 |  Researchers find 1,500+ exposed Hugging Face API tokens with write access | Credential exposure | Hugging Face platform (tokens tied to Meta, Google, Microsoft, VMware and others) | 1 |
| 2023-09-18 |  Microsoft AI research storage exposed through an overly permissive Azure SAS token | Credential exposure | Microsoft | 2 |
| 2023-05-01 |  Samsung restricts staff use of generative AI after sensitive data reached ChatGPT | Data exposure | Samsung Electronics | 2 |
| 2023-04 |  OpenAI's internal messaging systems breached in 2023, disclosed in 2024 | Data exposure | OpenAI | 1 |
| 2023-03-20 |  ChatGPT outage and data exposure from a redis-py client bug | Data exposure | ChatGPT | 1 |
| 2023-03-03 |  Meta's LLaMA model weights leak onto 4chan via BitTorrent | Data exposure | Meta LLaMA | 1 |
| 2022-09-12 |  Twitter users hijack Remoteli.io's GPT-3 job bot with prompt injection | Prompt injection | Remoteli.io Twitter bot (OpenAI GPT-3) | 1 |