ai-shield-audit
bởi laurentaia
Coding Agents & IDEsTóm tắt
Security audit engine for OpenClaw configurations.
Mô tả
Kỹ năng này làm gì
Security audit engine for OpenClaw configurations. Scans for vulnerabilities across 11 categories including secret leakage, over-privileged agents, network exposure, and misconfigurations. Outputs a structured JSON report with risk scores and remediation steps.
Covers 11 specific vulnerability categories in a single pass rather than requiring manual review across gateway auth, network settings, plugin config, and secret handling separately.
Khi nào nên dùng
- Checking a config for exposed API keys before sharing
- Hardening an OpenClaw instance before production deployment
- Reviewing subagent permissions for over-privilege
- Stripping secrets from a config file before publishing
- Running a pre-flight security check on a new OpenClaw setup
Quy trình ví dụ
User asks: audit my OpenClaw config for security issues
- 1
Load the config file at ~/.openclaw/openclaw.json
- 2
Run shield.js audit with --summary flag
- 3
Parse the returned JSON for risk_level and vulnerabilities array
- 4
Present prioritized findings with remediation steps
- 5
Offer to sanitize the config if any secrets are detected
Structured report with risk_level (e.g. HIGH), overall_score out of 100, list of vulnerabilities with descriptions, and safe_to_deploy flag
Cài đặt
clawhub install laurentaia/ai-shield-auditChạy lệnh CLI, hoặc dán prompt vào Claude Code, Cursor, Codex hay coding agent bạn đang dùng.