13 incidents

Risk: Credential and privilege abuse Control: Route to a person and opt-out Outcome: Not stated in sources

EffectiveNot effectiveNot effective (bypassed)Did not existNot stated in sourcesNot yet checkedAttacker's ownOrder: model · harness · application · target

Date	Incident	Risk	Controls mentioned in sources
2026-10-18	Vibe-coded South African website hacked six days after launch	Credential and privilege abuse	Not effectiveSecrets kept out of the model's context L3 Did not existChange review and separation of duties L3
2026-10-18	OpenAI research model accessed a cryptocurrency wallet without permission OpenAI	Credential and privilege abuse	Did not existHuman approval for high-impact actions L3
2026-10-15	Storm-0940 uses ARTEX agentic AI framework to automate cross-tenant password spraying and OAuth application abuse	Credential and privilege abuse	Did not existActivity logs L3 EffectiveStop control L2
2026-09-21	Malicious HEIF upload reached OpenAI internal GitHub	Credential and privilege abuse	EffectiveFlaw reporting and bug bounty L1
2026-09-16	Shai-Hulud worm spreads via hijacked AI coding assistant session	Credential and privilege abuse	Not effectiveLeast-privilege access for the agent L3 Did not existLive behaviour monitoring L3
2026-09-15	AEPD-confirmed autonomous AI agent data breach (Spain)	Credential and privilege abuse	Did not existNetwork segmentation L4 Did not existHuman approval for high-impact actions L3 Did not existRisk assessment per deployment L3
2026-07-17	Gemini lock-screen flaw lets strangers send messages from locked Android phones Google	Credential and privilege abuse	Not effectiveFail-closed defaults L2
2026-07-01	Claude helped hacker issue tickets to almost every US music festival Anthropic	Credential and privilege abuse	Did not existMulti-factor sign-in and password policy L4
2026-05-15	Unreleased OpenAI model used a leaked GitHub API key, attempted disposable-email signups, then fabricated earnings data during RL training OpenAI	Credential and privilege abuse	EffectiveLive behaviour monitoring L3
2026-04-20	Vercel breach via third-party AI tool Context.ai with broad Google Workspace OAuth access Context.ai	Credential and privilege abuse	EffectiveEncryption of stored data L3
2026-03-23	AI-assisted phishing campaign via Railway compromises hundreds of organizations	Credential and privilege abuse	Not effectivePhishing and fraud filters L4 Not effective (bypassed)Multi-factor sign-in and password policy L4
2026-02-20	hackerbot-claw AI bot actively exploits GitHub Actions workflows	Credential and privilege abuse	Did not existLeast-privilege access for the agent L3 EffectivePrompt-injection detection L2 EffectiveTool allow-list L2 Not effective (bypassed)Change review and separation of duties L3
2026-01-23	Clawdbot/Moltbot AI agent gateways exposed on the internet with secrets accessible Peter Steinberger	Credential and privilege abuse	EffectiveMulti-factor sign-in and password policy L4 Not effectiveFail-closed defaults L2 Not effectiveFail-closed defaults L2