OpenClaw Skills Can Leak Your API Keys
in Seconds.
Join 50 founders protecting their AI agent fleets
Fleet Shield scans, detects, and blocks malicious AI skills before they execute. Protect your agents, your data, and your infrastructure.
⚡ Free forever for your first scan. No credit card.
This skill attempts to read environment variables
Risk Score: 87/100 | Critical: Data Exfiltration Pattern Detected
This is what Fleet Shield caught in a popular OpenClaw skill last week.
What We Actually Do
Scan Every Skill
53-pattern detection for data exfiltration, credential theft, and malicious code.
Block High-Risk Code
Pro plans automatically block skills with risk scores above your threshold.
Keep Audit History
Full logs of every skill scanned, blocked, or allowed. Compliance ready.
Automatic Protection. Not Just Warnings.
Connect your OpenClaw runtime to Fleet Shield. Every skill is checked before execution. High-risk skills get blocked instantly.
async function beforeSkillRun(skillId, skillContent) {
const decision = await fleetShield.check(skillId, skillContent);
if (!decision.allowed) {
throw new Error(`Blocked: ${decision.reason}`);
}
}Simple Pricing
Start free. Upgrade when you need protection.
🔥 Founder pricing ends when we hit 50 paying users
Pro
Founder rate — 40% off forever
- Unlimited scans
- Auto-block high-risk
- 1 runtime connection
- Full history
Pro Annual
Early Protector Price — locked forever
- Everything in Pro
- Founder status
- Priority support
- Price locked for life
Agency plan available: $49/mo for 5 runtimes + team access.
Founding Member Pricing
We're opening 50 founding spots at 40% off forever. After that, pricing goes back to standard rates.
FAQ
Does this connect to my OpenClaw?
Yes. Pro plans get a runtime token. Add it to your OpenClaw config and every skill is checked before execution.
Do I need to install anything?
Just a single environment variable. No agents, no complex setup. Takes 2 minutes.
What happens if I ignore warnings?
With Pro, high-risk skills are blocked automatically. On Free, you get warnings but execution continues.
Is my code stored?
Skill content is analyzed in memory and discarded. We keep only the scan result and risk score for your audit log.