BLOG

Security Insights & Updates

Stay informed with the latest security research, product updates, and engineering deep-dives from the Panguard AI team.

SAFE-MCP Gets $12.5M. Here Is How ATR Fits Into the Stack.

Six tech giants funded SAFE-MCP to secure AI agent tool calls. ATR provides the detection layer that makes standards enforceable. Here is how they fit together.

Panguard AI Team7 min readMarch 27, 2026

Threat IntelligenceMarch 27, 2026

We Scanned 36,394 AI Agent Skills. 13.5% Have Security Risks.

The first full ecosystem scan of ClawHub is complete. 36,394 skills crawled, 9,676 with parseable content. 182 CRITICAL, 1,124 HIGH. 249 packages combine shell access, network calls, and filesystem operations. Here is everything we found.

Security Insights & Updates

SAFE-MCP Gets $12.5M. Here Is How ATR Fits Into the Stack.

We Scanned 36,394 AI Agent Skills. 13.5% Have Security Risks.

49% of npm MCP Packages Have Security Findings

The Control Plane Just Collapsed. 40 Years of Security Assumptions Are Gone.

Better AI Models Are More Hackable, Not Less

Clean for 15 Versions. Version 1.0.16 Stole Your Inbox.

The Email That Waited 6 Weeks to Steal Your Data

What Is ATR? The First Open Standard for AI Agent Security

OWASP Agentic Top 10: How 108 ATR Rules Cover All 10 Categories

Why Transparent 39.9% Recall Beats Opaque 98%

The Flywheel: Every Scan Makes Everyone Safer

Your AI Agent Can Read Your SSH Keys. OWASP Finally Noticed.

Skill Auditor npm Package: The Quick Start Guide

How ATR Protects Your AI Agent: A Real-World Guide

Why We Created ATR -- The Missing Detection Standard for AI Agents

ATR in Action: How Agent Threat Rules Caught a Tool Poisoning Attack

Contributing to ATR: How Security Researchers Can Shape AI Agent Security

Protect Your AI Agents: How to Audit Skills Before Installing

Panguard AI v0.3.0: Skill Auditor, MCP Tools & Public Release

Why 99% of Businesses Have Zero Security -- And What We're Doing About It

The Vibe Coding Security Gap: When AI Writes Code, Who Watches the Server?

90% of Threats Don't Need AI: Why We Built a Rules-First Security Engine

Why AI-Powered Threat Detection Outperforms Rule-Based Systems

The True Cost of a Data Breach for Small Businesses in 2026

Introducing Panguard Scan: Free 60-Second Security Audits

Honeypot Intelligence: How Panguard Trap Learns from Attackers

SOC 2 Compliance for Startups: A Practical Guide

Lateral Movement Detection: How Panguard Guard Stops Attackers

Building Security That Speaks Human: The Design Philosophy Behind Panguard Chat

Open Source Security: Why We Publish Our Scan Engine

Why AI Agents Need Their Own Sigma: The Case for ATR

The MCP Supply Chain Problem Nobody Is Talking About

How Collective Defense Closes the AI Security Gap for 400 Million Businesses

We Scanned 53,000 AI Agent Skills. Here Is What We Found.

30 CVEs in 60 Days: The MCP Attack Surface Is Exploding

ATR vs YARA vs Sigma: Detection Rules for AI Agents

How Cisco Ships ATR Rules in AI Defense

OWASP Agentic Top 10: Full Detection Coverage Mapping

Weekly Security Digest

Security Insights & Updates

SAFE-MCP Gets $12.5M. Here Is How ATR Fits Into the Stack.

We Scanned 36,394 AI Agent Skills. 13.5% Have Security Risks.

49% of npm MCP Packages Have Security Findings

The Control Plane Just Collapsed. 40 Years of Security Assumptions Are Gone.

Better AI Models Are More Hackable, Not Less

Clean for 15 Versions. Version 1.0.16 Stole Your Inbox.

The Email That Waited 6 Weeks to Steal Your Data

What Is ATR? The First Open Standard for AI Agent Security

OWASP Agentic Top 10: How 108 ATR Rules Cover All 10 Categories

Why Transparent 39.9% Recall Beats Opaque 98%

The Flywheel: Every Scan Makes Everyone Safer

Your AI Agent Can Read Your SSH Keys. OWASP Finally Noticed.

Skill Auditor npm Package: The Quick Start Guide

How ATR Protects Your AI Agent: A Real-World Guide

Why We Created ATR -- The Missing Detection Standard for AI Agents

ATR in Action: How Agent Threat Rules Caught a Tool Poisoning Attack

Contributing to ATR: How Security Researchers Can Shape AI Agent Security

Protect Your AI Agents: How to Audit Skills Before Installing

Panguard AI v0.3.0: Skill Auditor, MCP Tools & Public Release

Why 99% of Businesses Have Zero Security -- And What We're Doing About It

The Vibe Coding Security Gap: When AI Writes Code, Who Watches the Server?

90% of Threats Don't Need AI: Why We Built a Rules-First Security Engine

Why AI-Powered Threat Detection Outperforms Rule-Based Systems

The True Cost of a Data Breach for Small Businesses in 2026

Introducing Panguard Scan: Free 60-Second Security Audits

Honeypot Intelligence: How Panguard Trap Learns from Attackers

SOC 2 Compliance for Startups: A Practical Guide

Lateral Movement Detection: How Panguard Guard Stops Attackers

Building Security That Speaks Human: The Design Philosophy Behind Panguard Chat

Open Source Security: Why We Publish Our Scan Engine

Why AI Agents Need Their Own Sigma: The Case for ATR

The MCP Supply Chain Problem Nobody Is Talking About

How Collective Defense Closes the AI Security Gap for 400 Million Businesses

We Scanned 53,000 AI Agent Skills. Here Is What We Found.

30 CVEs in 60 Days: The MCP Attack Surface Is Exploding

ATR vs YARA vs Sigma: Detection Rules for AI Agents

How Cisco Ships ATR Rules in AI Defense

OWASP Agentic Top 10: Full Detection Coverage Mapping

Weekly Security Digest