Corrections
Every change made to a record's classification after it was published, with the reason. Reports come from the form on each record page; the classifier audit and the maintainer's own review find the rest. Dataset releases are frozen, so a correction applies to the live site and to the next release.
- Corrections
- 633
- Reports received
- 0
- Reports declined
- 0
- Reports open
- 0
Each correction is listed as its own row. Group bulk changes
| Date | Record | Change | Reason | Found by |
|---|---|---|---|---|
| 2026-10-10 | Recovering Encrypted LLM Reasoning Traces | Severity: high to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | How MCP Servers Can Expose Enterprise Secrets | Severity: high to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | Irregular Details How a Naming Error Let AI Models Attack a Real Company | Severity: high to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | Wiz Red Agent Finds Its Way Into Snowflake’s Internal Jira Due to an AI-Generated GitHub Copilot “Autofix” | Severity: high to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | Microsoft Copilot Personal Flaws Could Let One Click Exfiltrate Data From Connected Apps | Severity: high to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | OpenAI Overhauls Safety Protocols After Its AI Agents Went Rogue | Severity: high to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | Microsoft finally patches critical one-click Copilot vulnerability, almost eight months after learning of it | Severity: high to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | Snowflake flaw slips past AI checks, gets exploited by another AI | Severity: high to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | New Cryptographic Context Injection Attack Could Let Web Pages Steal Grok Chat Data | Severity: high to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | Critical flaw patched in popular JavaScript sandbox used in AI projects | Severity: high to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | More Incidents of AIs Going Rogue in Cybersecurity Challenges | Severity: high to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | Encrypted Prompts Bypass AI Safety Guardrails in Grok and Gemini | Severity: high to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | Marimo Notebook Flaw Could Run MCP Commands Before Cells Execute in Edit Mode | Severity: high to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | A Malicious Webpage Could Poison Your Local AI Model Behind NVIDIA NemoClaw | Severity: high to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | Finding Nemo(Claw): Networking Issue Allows for LLM Poisoning in OpenClaw | Severity: high to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | AnonyMousKIT PhaaS uses voice AI agents to phish iPhone passcodes | Severity: high to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | Claude Opus 4.6 Bypasses Gym Booking Limit, Cancels Other Users' Reservations in Tests | Severity: high to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | VMs won't contain cyber-capable agents | Severity: high to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | NemoClaw’s AI can be poisoned through a browser tab | Severity: high to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | What We Still Don’t Know About OpenAI’s Hugging Face Hack | Severity: high to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | OpenAI’s rogue AI model incident was worse than we thought | Severity: high to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | Breaking Claude Code Opus 5 Auto Mode | Severity: high to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | OpenAI Agents Coordinated via Makeshift Message Board Ahead of Hugging Face Hack | Severity: high to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | Amazon Kiro Prompt Injection Can Exfiltrate Sensitive Data Through Kiro Powers | Severity: high to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | Inside 90 days of attacks on AI infrastructure | Severity: high to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | OpenAI Says Reward Hacking Drove AI Agents to Exploit Zero-Days and Breach Hugging Face | Severity: critical to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | Nearly 700 rogue AI agents coordinated in the Hugging Face attack | Severity: critical to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | Breaking Claude Code Opus 5 Auto Mode | Severity: high to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | OpenAI Agents Exploited Linux Kernel Flaw on Company’s Own Systems | Severity: high to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | GPUThor hardware attack can root Nvidia GPU systems | Severity: high to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | Hundreds of OpenAI Agents Invaded Hugging Face Servers | Severity: high to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | ServiceNow patches three maximum severity flaws that could put enterprise data at risk | Severity: critical to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | Anthropic warns infostealer malware is hijacking Claude sessions to drain usage | Severity: high to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | Anthropic Warns Claude Users of Infostealer Malware Infections | Severity: high to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | ⚡ Weekly Recap: Chinese Spy Proxy, AI Agents Go Off-Task, Router Backdoors and More | Severity: high to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | Anthropic Users Hit by Infostealer Attacks, Session Thefts | Severity: high to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | Attackers Steal METR API Key and Consume AI Credits Worth About $600,000 | Severity: high to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | Path to Astra: critical capabilities and frontier safeguards | Severity: high to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | Financially Motivated Threat Actor BREEZE COMET Targets Brazil | Severity: high to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | Critical Langflow flaw exploited to steal OpenAI and AWS keys | Severity: critical to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | AI Model Evaluator METR Hit by Credential Theft, Probing | Severity: high to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | When the patch tsunami meets the maintenance window | Severity: high to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | An AI-Assisted Cyber Attack: Inside a Unit 42 Investigation | Severity: high to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | Anthropic Details Response to Security Incidents, Unveils Enterprise Safeguards | Severity: high to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | Malicious .git Configs Can Make Claude, Codex, Cursor, and Other AI Agents Run Attacker Code | Severity: high to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | AI agents help compress ransomware intrusion to under 10 hours, raising stakes for CISOs | Severity: high to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | Anthropic's distillation battle turns to the dark web as China concerns swell | Severity: high to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | AI Coding Agents Are Installing Unknown/Untrusted Code on Corporate Networks | Severity: high to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | OpenAI agents hijacked German website before Hugging Face hack, report claims | Severity: high to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |
| 2026-10-10 | Thousands of OpenAI Agents Quietly Turned an Abandoned Wiki Into Their Coordination Channel | Severity: high to medium | Severity capped at medium. High and critical measure technical exploitability and are reserved for vulnerabilities and incidents; this record is a news report. | Classifier audit |