| name | openclaw-guard |
| description | Universal security control for OpenClaw agents. Prevents overspending (token/money budgets), unauthorized posting, file deletion, and protects itself from tampering. No telemetry by default; webhook alerts only if configured. |
| version | 1.0.0 |
| metadata | {"openclaw":{"requires":{"bins":["node"]},"install":[{"kind":"node","package":"openclaw-guard","bins":["openclaw-guard"]}],"emoji":"🛡️","homepage":"https://github.com/myaist/openclaw-guard","os":["macos","linux","windows"]}} |
OpenClaw Guard 🛡️
Four core protections, ready out of the box:
| Protection | Description |
|---|
| 💰 No overspending | Set token/call/money limits. Free use within limits; auto-block when exceeded. |
| 📢 No unauthorized posting | POST/email denied by default; must be explicitly allowed in policy. |
| 🗑️ No file deletion | All delete operations blocked globally. |
| 🔒 Tamper-proof | Plugin files are completely locked; agents cannot modify them. |
Install
npm install -g openclaw-guard
Register as an OpenClaw hook
Edit ~/.openclaw/settings.json:
{
"hooks": {
"PreToolUse": [
{
"matcher": "*",
"hooks": [
{
"type": "command",
"command": "openclaw-guard-hook",
"timeout": 10
}
]
}
]
}
}
Minimal config
Create security-config.yaml:
redis:
url: "redis://localhost:6379"
storage:
auditLog:
backend: file
path: "./logs/audit"
policy:
path: "./policies"
defaultEffect: deny
alerting:
channels:
- type: stdout
failsafe:
mode: fail-closed
Control LLM spending
version: "1"
scope: agent
target: "llm-agent"
defaultEffect: allow
budget:
daily:
tokens: 1000000
calls: 500
singleOp:
tokens: 100000
rateLimits:
- action: "call:llm"
window: 1m
maxCount: 30
View audit log
openclaw-guard audit --limit 50
openclaw-guard audit --decision deny
Links
