ai-agents

The 9-Second Disaster: What a Rogue AI Coding Agent Teaches Us About Production Access

John Z Black Apr 27, 2026

AI Security #ai-agents #cursor #claude #developer-security #vibe-coding

A Claude-powered agent deleted an entire production database in 9 seconds. Here's why it happened and what it means for anyone using AI coding tools.

Docker's Authorization Bypass Is Back. It's Been Broken for a Decade.

John Z Black Apr 8, 2026

Vulnerabilities & Patching #docker #container-security #authz-bypass #cve-2026-34040 #ai-agents #sandbox-escape

Pad an HTTP request past 1MB and Docker's AuthZ plugins see nothing. CVE-2026-34040 has been sitting in Docker Engine since 2016, and researchers showed AI agents can be tricked into exploiting it.

OpenClaw's Sixth Pairing Bug in Six Weeks Is a Full Admin Takeover

John Z Black Apr 4, 2026

Vulnerabilities & Patching #openclaw #cve-2026-33579 #privilege-escalation #ai-agents #supply-chain

CVE-2026-33579 lets anyone with the lowest access level become full admin on OpenClaw. It's the sixth pairing CVE in six weeks, and 63% of instances run without auth.

AI Governance Is an Implementation Problem Now, Not a Policy Project

John Z Black Mar 19, 2026

AI Security #ai-governance #ai-agents #least-privilege #data-locality #enforcement #cloudflare #unit42

Unit 42 on agent risk, Cloudflare on data-locality controls, and the ICML enforcement controversy all point to the same thing: governance only counts when it's technically enforceable and organizationally defended.

The Week Trust Kept Breaking

John Z Black Mar 15, 2026

Weekly Recap #the-week-in-cyber #iran-cyber #supply-chain #ai-agents #patch-management #ransomware #privacy #cisa

Iranian wipers, poisoned dev tools, AI agents as attack surfaces, patches that never stopped coming, and a ransomware negotiator working for the bad guys. Trust fell apart in every direction this week.

AI Agents Have a Security Problem — and It's Not Science Fiction Anymore

John Z Black Mar 15, 2026

AI Security #ai-agents #prompt-injection #llm-security #agentic-ai #enterprise-ai #owasp-llm

AI agents aren't chatbots. They act, execute, and chain decisions on their own. And the security model for most deployments? Basically nonexistent.

AI Agents Have an Infrastructure Problem — and Researchers Just Proved It

John Z Black Mar 11, 2026

AI Security #ai-security #mcp #red-team #llm #privacy #ai-agents

MCP protocol flaws, a 38-researcher red team exercise, and LLM-powered deanonymization all landed the same week. AI agent security isn't a future problem. It's a right now problem.