# AI Agent Incident Report — November 2026

*314 publicly reported incidents involving autonomous AI agents this month · 0 high-risk · tracked at https://watch.cloudaitradingagent.com*

## Most notable incidents

- **The Harness Is the Hard Part of Agentic Pentesting** (risk 69) — An AppSec team can point an AI agent running on a general-purpose model at its own web apps with little setup. Sure, the model writes the attacks. But everything that makes those attacks safe to run a. [Report](https://thehackernews.com/2026/10/the-harness-is-hard-part-of-agentic.html)
- **N-able Takes N-central Offline as Researcher Drops Over 30 Zero-Days** (risk 66) — Attackers could exploit the flaws for remote code execution, credential theft, and administrator account creation. The post N-able Takes N-central Offline as Researcher Drops Over 30 Zero-Days appeare. [Report](https://www.securityweek.com/n-able-takes-n-central-offline-as-researcher-drops-over-30-zero-days/)
- **Hackers use AntiFox AI agents in AWS credential-theft attacks** (risk 62) — Attackers are using Alibaba Cloud to host self-managed instances of AntiFox, an open-source agentic penetration testing platform, to target web services and AWS environments. [...]. [Report](https://www.bleepingcomputer.com/news/security/hackers-use-antifox-ai-agents-in-aws-credential-theft-attacks/)
- **AntiFox AI Pentesting Platform Linked to Attacks That Accessed AWS Production Secrets** (risk 58) — Cybersecurity researchers have disclosed details of a campaign in which an emerging agentic penetration testing platform called AntiFox has been used to steal Amazon Web Services (AWS) credentials and. [Report](https://thehackernews.com/2026/11/antifox-ai-pentesting-platform-linked.html)
- **Chainalysis Confirms Intrusion Into Internal Tools and Exposure of Customer Identity Data** (risk 58) — American blockchain analysis firm Chainalysis said it was the target of a cybersecurity incident in which an external actor accessed non-sensitive customer data. However, the company noted that the br. [Report](https://thehackernews.com/2026/10/chainalysis-confirms-intrusion-into.html)
- **In-the-Wild Attack Abuses AI Agents' Trust in Markdown** (risk 56) — A new AI agent prompt injection technique exploits how coding assistants consume local instruction files, blurring the line between text documents and executable actions. [Report](https://www.darkreading.com/application-security/in-the-wild-attack-ai-agents-trust-markdown)
- **Google Finds Hackers Using AI Agents to Build and Run Exploits From Software Patches** (risk 55) — Google on Monday revealed that adversarial use of artificial intelligence (AI) by threat actors has gone beyond manual prompting to deploying autonomous multi-agent offensive frameworks like ARTEX, CA. [Report](https://thehackernews.com/2026/11/google-finds-hackers-using-ai-agents-to.html)

## By the numbers

- **Most-affected organizations:** Unknown (314)
- **Top origin countries:** Unknown (314)
- **Common techniques:** T1190 Exploit Public-Facing Application (136), T1078 Valid Accounts / Credential Abuse (83), T1052 Exfiltration (46), T1059 Command & Scripting Execution (22), T1105 Ingress Tool Transfer (14), T1021 Lateral Movement (7)
- **Confirmed incidents:** 0
- **All-time tracked:** 364

## What this tells us

Agent incidents keep clustering around the same failure modes: instruction-following taken too literally, credential access without scoping, and actions taken without human gates. If you're building agents that read untrusted content, put controls in front of what they read and what they can do.

*Data: https://watch.cloudaitradingagent.com — public, attributed incidents only. Updated continuously.*