🛡️ ClawShield Pro
Scan OpenClaw/Codex skills for security risks
What We Detect
💉
Prompt Injection
Detect instruction override patterns that could hijack your AI agent's behavior.
📤
Data Exfiltration
Find unauthorized network calls and hardcoded URLs that leak sensitive data.
⚠️
Dangerous Commands
Flag rm -rf, curl | bash, and other risky shell commands.
🔑
Secret Exposure
Detect hardcoded API keys, tokens, and credentials in your code.
📦
Dependency Audit
Check npm/pip dependencies for known CVEs and vulnerabilities.
🏷️
Trust Badge
Get a verifiable badge to show your skill has been security-audited.