SkillAgentSearch skills...

ClawShield

Security preflight and guardrails for OpenClaw/Moltbot

Install / Use

/learn @kappa9999/ClawShield
About this skill

Quality Score

0/100

Supported Platforms

Universal

README

<div align="center"> <img src="assets/logo.svg" width="140" alt="ClawShield logo" /> <h1>ClawShield</h1> <p>Security preflight and guardrails for OpenClaw/Moltbot</p> <p> <img alt="CI" src="https://github.com/kappa9999/ClawShield/actions/workflows/ci.yml/badge.svg" /> </p> </div>

What is ClawShield?

ClawShield is a safety helper for OpenClaw/Moltbot users. It checks your config for risky settings, warns you if your gateway is exposed, and helps you keep skills from being tampered with.

ClawShield does not change your config unless you explicitly opt in with clawshield apply safe --write.

Why this matters (plain English)

OpenClaw can run powerful tools. That is great, but it also means a mistake in config can expose your control panel or let the wrong agent use dangerous tools. ClawShield helps you catch those issues early.

Common risks it helps prevent:

  • Gateway bound to the whole network without auth.
  • Group chats where anyone can trigger actions without a mention.
  • Non-main agents running without a sandbox.
  • Skills modified after install (supply-chain risk).

Quick install (macOS/Linux)

curl -fsSL https://raw.githubusercontent.com/kappa9999/ClawShield/main/scripts/install.sh | bash

Install (Homebrew)

brew tap kappa9999/clawshield
brew install clawshield

Install (local dev)

npm install
node ./src/cli.js --help

Install (global)

npm install -g clawshield
clawshield --help

How to use (step-by-step)

1) Run a basic audit

clawshield audit

You will see a report with PASS/WARN/FAIL items and suggested fixes.

2) Check if your gateway is exposed

clawshield exposure

If it shows EXPOSED listeners, your control port is reachable outside of localhost and should be locked down.

3) Generate a safe profile snippet

clawshield profile safe

This prints a JSON snippet you can merge into openclaw.json.

4) Apply a safe profile (opt-in)

clawshield apply safe
clawshield apply safe --write --token "YOUR_TOKEN"
  • The first command shows a dry-run diff.
  • The second writes the changes and creates a .bak-* backup.

5) Lock your skills (tamper detection)

clawshield lock
clawshield verify
  • lock creates a fingerprint of installed skills.
  • verify tells you if anything changed or is missing.

6) Watch exposure in the background (macOS)

clawshield watch --interval 30

Or generate a LaunchAgent:

clawshield launchagent --bin /usr/local/bin/clawshield --interval 30 --write ~/Library/LaunchAgents/com.clawshield.watch.plist
launchctl load ~/Library/LaunchAgents/com.clawshield.watch.plist

Notes

  • Default config path: ~/.openclaw/openclaw.json (override with OPENCLAW_CONFIG_PATH).
  • Default workspace: ~/.openclaw/workspace.
  • The audit is read-only; only apply safe --write changes your config.

Troubleshooting

  • If clawshield is not found, ensure npm global bin is on your PATH.
  • If the audit cannot read your config, pass --config /path/to/openclaw.json.
  • On macOS, the exposure check uses lsof. Make sure it is available.

Roadmap

  • ClawdHub provenance integration (auto-verify installed skill sources).
  • Optional notifications (desktop/Slack/Discord).
  • Guided remediation mode.
View on GitHub
GitHub Stars19
CategoryDevelopment
Updated20d ago
Forks2

Languages

JavaScript

Security Score

90/100

Audited on Mar 10, 2026

No findings