4 min read
How to Check if an OpenClaw Skill is Safe Before Installing It
Before you install any OpenClaw skill, make sure it's safe. A practical checklist for evaluating AI agent skill security — and when to use a scanner.
Read article →Blog
Security insights, research, and updates from the ClawDefend team.
5 min read
The Hidden Security Risks in AI Agent Skills (And How to Scan Them)
AI agents are only as safe as the skills you give them. The four most common vulnerabilities ClawDefend finds — and what to look for in your own skill directory.
Read article →4 min read
How We Measured That 70%+ Stat — And Why We Tightened It
We found a flaw in our own scanner: it was flagging the absence of bad behavior as a finding. Here's how we fixed it and what the real number is.
Read article →4 min read
How ClawDefend Works: Static Analysis, AST Parsing, and AI Intent Detection
A technical deep-dive into our scanning pipeline — from downloading the ZIP to generating a line-level security report.
Read article →3 min read
AI Agent Marketplaces Are the npm of 2026 — And Nobody's Watching
Before the npm supply chain attacks, everyone trusted packages without question. AI agent skill marketplaces are at that exact inflection point.
Read article →3 min read
We Scanned 851 AI Agent Skills — Here's What We Found
Over 70% of skills on ClawHub had medium or higher severity security findings. Here's a breakdown of the most common vulnerabilities.
Read article →