01
Drop in your skills
Upload one ZIP or a whole toolkit. We unpack it in memory—not on a live filesystem.
AI SKILL SECURITY ANALYZER
Powerful skills. Safer agents.
Your AI skills can do a lot. Make sure it's what you intended.
Uncover hidden threats before you give them access.
Let's look inside.
Upload your skills. Know what you're giving access to.
Add a second perspective. Off by default for privacy.
Just exploring?
Your code is never executedUploads aren't savedNo account needed
Built for the open Agent Skills ecosystem
Claude CodeOpenClawCursorCodex+ any SKILL.md package
LESS GUESSWORK. MORE CONFIDENCE.
From unknown to understood.
02
Look beneath the surface
Inspect instructions, scripts, and supporting text for risky patterns. Add AI review for more context.
03
Make an informed call
See what was found, where it lives, and what to do next. Download a report you can share.
BUILT ON REAL-WORLD THREAT RESEARCH
More than a suspicious line of code.
Understand the attack surface—and exactly where automated analysis stops.
Your skills stay yours. Skill Guard processes uploads in server memory and does not save archives or reports. Reports remain in this page until you leave or clear them; download a copy to keep it. Optional AI review sends best-effort redacted text to an AI provider through Vercel AI Gateway, where provider retention policies apply. Remove secrets before uploading. Always pair automated analysis with human review and least-privilege access.