Complete guide to integrating MoltGuard with your AI agent.
Connect your Clawdbot agent to MoltGuard in under a minute:
| Option | Type | Default | Description |
|---|---|---|---|
| token | string | required | Your MoltGuard API token |
| url | string | http://guard.moltnet.ai | MoltGuard server URL |
| agentName | string | auto | Display name in dashboard |
| logThoughts | boolean | true | Send reasoning to Mind Graph |
| gateHighRisk | boolean | true | Require approval for high-risk actions |
| gateTools | string[] | ["exec", "message", "write"] | Tools requiring approval |
| pollCommands | boolean | true | Enable remote control |
| pollIntervalMs | number | 5000 | Command poll interval |
📦 Plugin source: github.com/moltnet/guard/plugin
Base URL: http://guard.moltnet.ai/api
Scan a skill for security issues. Provide either a URL or raw content.
| Parameter | Type | Description |
|---|---|---|
| url | string | URL of the skill file to scan |
| content | string | Raw skill content to scan |
| name | string | Optional skill name |
Get recent scan history.
Get an SVG badge showing the scan result. Use in README files.
Log an action from your agent. Use status="pending" to require approval.
| Parameter | Type | Description |
|---|---|---|
| agent | string | Agent identifier |
| type | string | Action type (e.g. email.send, file.write) |
| description | string | Human-readable description |
| risk | string | low | medium | high | critical |
| status | string | executed | pending |
| details | object | Additional structured data |
| costUsd | number | API cost in USD |
| tokensIn | number | Input tokens used |
| tokensOut | number | Output tokens used |
| reversible | boolean | Can this action be undone? |
| undoAction | object | Instructions for undoing |
| callbackUrl | string | Webhook URL for decision notifications |
Poll for the decision on a pending action. Returns status, decidedAt, decidedBy, rejectReason.
Get all pending actions awaiting approval.
Approve a pending action.
Reject a pending action. Optionally include reason in body.
Mark a reversible action as undone. Returns undoAction if provided.
Get aggregate statistics: action counts, costs, risk breakdown, top agents.
Get hourly activity data for charts.
Health check endpoint.
Register a new agent. Returns API key for authentication.
List all registered agents.
Start a new agent session. Returns sessionId.
Log a thought/reasoning step. Used to build the mind graph.
| Parameter | Type | Description |
|---|---|---|
| sessionId | string | Session ID |
| agent | string | Agent name |
| type | string | reasoning | decision | planning | observation | reflection |
| title | string | Brief title |
| content | string | Detailed content |
| parentId | string | Parent trace ID (for nesting) |
Get mind graph data (nodes and edges) for visualization.
Send a control command to an agent (pause, resume, stop).
Get pending commands for an agent (agent polls this).
Acknowledge command execution.