Skip to content

CicadaRelay/openclaw-hardener

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

1 Commit
 
 
 
 
 
 

openclaw-hardener

OpenClaw self-check & auto-hardening skill. Audits your instance for the exact vulnerabilities tracked by declawed.io — 230,000+ exposed instances worldwide. Don't be one of them.

12-Point Security Audit

# Check Severity Auto-Fix
1 Network Exposure (0.0.0.0 binding) CRITICAL Yes
2 Empty/Missing Authentication CRITICAL Yes
3 Favicon Fingerprint (declawed.io detection) HIGH Yes
4 CVE-2026-25253 Gateway Auth Bypass (CVSS 9.8) CRITICAL Yes
5 Docker Socket Escape CRITICAL Yes
6 API Key Exposure in Config/JS HIGH Yes
7 MCP Server Exposure HIGH Yes
8 CVE-2026-25157 SSH Injection (macOS) HIGH Partial
9 Bot/Messaging Channel on Exposed Instance MEDIUM Yes
10 Running as Root HIGH Yes
11 Missing Firewall Rules HIGH Yes
12 Malicious Skills (ClawHavoc Campaign) CRITICAL Quarantine

Install

# Copy to your OpenClaw skills directory
cp -r openclaw-hardener/ /path/to/openclaw/skills/
# Or symlink
ln -s $(pwd)/openclaw-hardener /path/to/openclaw/skills/openclaw-hardener

For Claude Code / OpenClaw users:

cp -r openclaw-hardener/ ~/.openclaw/skills/

Usage

Once installed, the AI agent can invoke the skill:

  1. list_skills — will show openclaw-hardener in the list
  2. read_skill("openclaw-hardener") — loads the full audit checklist
  3. Agent runs all 12 checks, outputs a scored report, and offers auto-fix

Scoring

Score Grade Action
95-100 A Hardened
85-95 B Acceptable
70-85 C Needs improvement
50-70 D Significant risk
< 50 F Immediate action required

Background

On 2026-03-09, we discovered our own OpenClaw instance (favicon MD5: f58854f6450618729679ad33622bebaf) was publicly exposed and likely already indexed by declawed.io's threat intelligence platform. We built this skill to prevent others from making the same mistake.

References

  • declawed.io — SecurityScorecard STRIKE Team threat intelligence
  • CVE-2026-25253 — Gateway auth bypass (CVSS 9.8)
  • CVE-2026-24763 — Docker sandbox escape (CVSS 8.8)
  • CVE-2026-25157 — SSH command injection (CVSS 7.8)

License

MIT

About

Self-check and hardening tools for OpenClaw deployments before public exposure

Resources

License

Code of conduct

Contributing

Security policy

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors