Security Audit

Comprehensive security auditing for Clawdbot deployments. Scans for exposed credentials, open ports, weak configs, and vulnerabilities. Auto-fix mode included.

Install only if you want a local Clawdbot security audit. Run the normal audit first, treat reports as sensitive because they may reveal paths and security posture, and use --fix only after reviewing backups because it can change file permissions and create a .gitignore file.

Type: OpenClaw Skill Name: security-audit Version: 1.0.0 The OpenClaw AgentSkills skill bundle is designed for comprehensive security auditing of Clawdbot deployments. The `SKILL.md` instructions clearly outline its purpose and usage, including options for scanning credentials, ports, configurations, and permissions, as well as an auto-fix mode. The core script, `scripts/audit.cjs`, uses `child_process.execSync` to run system commands like `ss` or `netstat` for port scanning and `git log` for commit history, which are legitimate actions for a security audit tool. It also uses `fs` module functions to read files and scan for sensitive patterns (e.g., API keys, passwords) within the defined `CLAWDBOT_DIR` (`/root/clawd`). Crucially, while it identifies sensitive data, there is no evidence of data exfiltration to external endpoints. The auto-fix functionality (`fs.chmodSync`, `fs.writeFileSync`) is used to harden security by setting restrictive file permissions and creating a `.gitignore` file. No prompt injection attempts against the agent, obfuscation, or other malicious indicators were found; the behavior is entirely aligned with its stated security auditing purpose.