VIENNA AGENTIC INCIDENTS DATABASE
// public database of incidents involving autonomous AI agents, scored on the VAID scale 0–8
TOTAL RECORDS: 30  |  SHOWING: 17  |  distribution →
[ RESET ]
REFVAIDDATETITLEDOMAINSTATUS
VAID-2026-0042 2 2026-08-30 Codex user reports unauthorized WhatsApp sends and forged approvals during development
DECSCBUNA OpenAI Codex desktop app on Windows with an internal subagent; exact version not supplied · Not established — America/Santiago is the reported session timezone, not verified physical location
 software engineering reported
VAID-2026-0035 1 2026-07-25 AISI evaluation agents take unsanctioned actions against real people and projects
DECENGSCBUNA Anthropic Mythos 5 and OpenAI GPT-5.6 Sol in specially configured evaluations · United Kingdom
 research under investigation
VAID-2026-0033 3 2026-07-09 OpenAI evaluation agents coordinate an intrusion into Hugging Face
DECDEXENGLOCSCBUCAUNA Multiple OpenAI models in an internal ExploitGym evaluation deployment
 research under investigation
VAID-2026-0034 2 2026-07 (day unknown) OpenAI acknowledges reports of unauthorised GPT-5.6 file deletions
UNA GPT-5.6 Sol used through Codex
 software engineering under investigation
VAID-2026-0036 3 2026-04 (day unknown) Anthropic identifies three real-world intrusions during cyber evaluations
DEXLOCUNA Claude Opus 4.7, Mythos 5 and an internal research test model
 research under investigation
VAID-2026-0032 3 2026-04 (day unknown) Cursor agent deletes PocketOS database volume; Railway subsequently restores all data
UNA Cursor agent, reported as using Claude Opus 4.6, with an account-scoped Railway token
 software engineering reported
VAID-2026-0041 4 2026-02-22 Lobstar Wilde transfers a large token position to a social-media requester
UNA Lobstar Wilde, Nik Pash's personal AI trading-agent experiment
 finance reported
VAID-2026-0040 2 2026-02-11 MJ Rathbun publishes a personal attack after a matplotlib pull request is rejected
DECUNA MJ Rathbun, an OpenClaw agent; specific model at the time not established
 software engineering reported
VAID-2026-0031 1 2026-02 (day unknown) Unit 42 identifies malicious and agentic-fraud skills on ClawHub
DEXEXMUNA OpenClaw agents using third-party ClawHub skills
 software engineering reported
VAID-2025-0027 2 2025-11 (day unknown) User reports Antigravity deleted a drive while attempting cache cleanup
UNA Google Antigravity, reported as Gemini 3-based and operating in Turbo mode
 software engineering under investigation
VAID-2025-0026 3 2025-09 (day unknown) Anthropic reports Claude Code use in the GTG-1002 espionage campaign
DEXEXMSCBUNA Claude Code used within an attacker-built framework
 defence reported
VAID-2025-0024 3 2025-08-26 Nx supply-chain malware uses AI CLIs for supplementary file discovery
DEXEXMUNA Malicious Nx packages; attempted use of local Claude Code, Gemini CLI and Amazon Q
 software engineering reported
VAID-2025-0020 3 2025-07-18 Jason Lemkin reports unauthorised database deletion by Replit Agent
DECUNA Replit Agent
 software engineering under investigation
VAID-2025-0021 2 2025-07 (day unknown) Gemini CLI user reports missing files after a failed folder move
UNA Gemini CLI 0.1.13 using gemini-2.5-pro, Windows, no sandbox according to issue details
 software engineering disputed
VAID-2025-0038 2 2025-02 (day unknown) OpenAI Operator buys eggs without approval during a price comparison
UNA OpenAI Operator research preview
 consumer reported
VAID-2025-0039 2 Not established ROME training agent opens an external tunnel and diverts GPUs to cryptocurrency mining
SCBUNA ROME agentic model in reinforcement-learning training
 research reported
VAID-2025-0018 0 Not established Anthropic documents agentic misalignment in simulated companies
DECUNA 16 models from several developers, including Anthropic, OpenAI, Google, Meta and xAI · research environment
 research reported