The Trusted Security Hub
for OpenClaw

OpenClaw itself is an open-source tool with strong security foundations. However, the skill ecosystem on ClawHub introduces risks. Malicious skills can steal credentials, execute code, and exfiltrate data. Use verified skills, enable sandbox mode, and follow our security guides to stay safe.

Malicious skills on ClawHub use several attack vectors: typosquatting (naming skills similarly to popular ones), data exfiltration (stealing .env files, SSH keys, API tokens), reverse shells (backdoor connections), and crypto miners. New malicious skills continue to appear, so always verify before installing.

Like any software, OpenClaw may have vulnerabilities in its skill loader and execution engine. These can allow remote code execution through crafted skill manifests. Always keep OpenClaw updated to the latest version and use sandbox mode to mitigate potential vulnerabilities.

The Skill Verifier runs entirely in your browser — no data is sent to any server. It checks skill names against our database of known malicious skills, compares with verified safe skills, and runs heuristic analysis on skill content looking for suspicious patterns like network exfiltration, reverse shells, and credential access.

Verified skills have been manually audited by our security team. We review the source code, check permissions, test for malicious behavior, verify the publisher identity, and monitor for changes. Verified skills receive a trust score of 89 or higher.

Report malicious skills through ClawHub's reporting mechanism and submit details to our team via the UseClawPro GitHub repository. Include the skill name, URL, and any suspicious behavior you observed. We'll investigate and add confirmed threats to our database.

Yes. OpenClaw works without any third-party skills. You can use it with only built-in capabilities, which are safe and audited. If you need extended functionality, only install skills from our verified catalog or skills you've personally reviewed.

We update the database continuously as new threats are discovered. The database is bundled with the site — when you use the verifier, you're checking against the latest known threats at the time of our last deployment.