The trust layer for AI agents

Independent safety grades for the AI agents you run.

Rate a skill before you install it. Guard the deployment you already run. Graded A to F against the standards your auditors recognize, with the evidence for every finding.

Check a skill, free Audit a deployment →

Graded against OWASP · MITRE ATLAS · NIST AI RMF · EU AI Act

Before you install it

Rate a skill

Paste a skill, get an A to F grade with the evidence and the fix.

Free to check. Certify for $79. Monitor for $29/mo.

Check a skill

While you run it

Guard a deployment

Connect a repo read-only and probe your live agent. One graded report on your whole agent estate, with a prioritized fix list.

From $500 once, or $199/mo to stay covered.

Audit a deployment

Trust, earned

We publish our own accuracy

Agreement with human review, plus our false-positive and false-negative rates, so you can judge the rater, not just the rating. See the methodology →

Pricing

Rate a skill for free. Guard a deployment from $500.

No credits. No metered bills. Payments handled by Creem.

Rate

Free check $0 · Certification $79 once · Monitoring $29/mo

See Rate pricing →
Guard

Posture Report from $500 once · Subscription from $199/mo

See Guard pricing →

See all pricing

Who it's for

If your AI touches your data, this is for you

Solo builders

One bad skill is your reputation. Rate it and earn a badge that wins installs.

Teams & SMBs

You run agents wired into email, files and payments. Guard the deployment →

Enterprise

Independent evidence for your EU AI Act, ISO 42001, and SOC 2 duties. For teams →

Get the State of AI Skill Security report

What we're measuring across the skills people actually run — the common failure modes, by the numbers. One email when it's out. No spam, unsubscribe anytime.