A lightweight security gateway that intercepts AI tool calls to enforce safety policies and human-in-the-loop approvals.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install authensor-gateway
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install authensor-gateway using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Authensor Gateway?
Authensor Gateway serves as a critical security layer for developers using Openclaw Skills. It functions as a fail-safe policy gate that monitors every tool call generated by an AI agent before execution. By evaluating actions against a predefined security policy, it ensures that low-risk operations proceed automatically while high-risk or potentially destructive actions are paused for human intervention. This skill is specifically designed to protect local environments from malicious or accidental commands without compromising the agent's productivity.
The skill prioritizes privacy by utilizing a metadata-only approach. Only the action type and a redacted resource identifier are sent to the control plane, ensuring that your source code, conversation history, and sensitive API keys never leave your local machine. It provides a robust defense mechanism when running third-party Openclaw Skills that might otherwise have broad permissions to modify the filesystem or access the network.
Authensor Gateway Use Cases
- Safely executing Openclaw Skills sourced from public marketplaces or third parties
- Preventing accidental execution of destructive commands like file deletions or database drops
- Implementing human-in-the-loop oversight for AI-driven code modifications and deployments
- Generating a comprehensive audit trail of agent activities for regulatory and compliance requirements
- Protecting sensitive directories such as .ssh and .env from unauthorized agent access
How Authensor Gateway Works
- The AI agent identifies a tool it needs to use and classifies the action type based on the defined protocol.
- The agent redacts any sensitive information, such as credentials or tokens, from the command or URL.
- A POST request is sent to the Authensor control plane containing the action metadata and the tool name.
- The control plane evaluates the request against your policy and returns a decision: allow, deny, or require approval.
- For actions requiring approval, the agent pauses and polls for a status update while the user authorizes the action via CLI or web.
- The tool is only executed if the final decision is allow or approved; otherwise, the agent informs the user of the restriction.
Authensor Gateway Setup
To activate the gateway for your Openclaw Skills, you must configure the environment variables in your local configuration file. Follow these steps:
- Obtain your API key from the Authensor platform.
- Add the following configuration to your
~/.openclaw/openclaw.jsonfile:
{
"skills": {
"entries": {
"authensor-gateway": {
"enabled": true,
"env": {
"CONTROL_PLANE_URL": "https://authensor-control-plane.onrender.com",
"AUTHENSOR_API_KEY": "your_authensor_api_key"
}
}
}
}
}
- Start a new session to load the updated security instructions.
Authensor Gateway Data Schema & Taxonomy
The skill utilizes a strict classification and redaction schema to manage tool metadata for Openclaw Skills. Data is organized into the following categories:
| Action Type | Description | Resource Redaction |
|---|---|---|
| safe.read | Reading non-sensitive files | Full path sent |
| filesystem.write | Writing or editing files | Full path sent |
| code.exec | Shell commands | Credentials/tokens stripped |
| secrets.access | Accessing .env or .ssh | Path blocked or flagged |
| network.http | Outbound web requests | Query parameters removed |
| dangerous.delete | File or directory removal | Full command flagged |
All interactions generate a unique receiptId for audit tracking.
Authensor Gateway Advanced Features
- Dual-layer enforcement combining system prompt instructions with a deterministic shell-level hook
- Fail-closed architecture that blocks all tool execution if the control plane is unreachable
- Sophisticated regex-based redaction of environment variables and inline tokens from bash commands
- Real-time polling mechanism for human-in-the-loop approvals within the agent workflow
- Support for custom policy definitions to granularly control Openclaw Skills permissions
SKILL.md
Loading
Related Openclaw Skills
ETHSkills
A dynamic knowledge base providing AI agents with real-time, accurate Ethereum documentation to bridge the gap between stale training data and modern blockchain production.
Airtable Participants
A professional integration for querying retreat participant data and managing contact segments directly from Airtable.
Travel Procurement
An intelligent procurement skill for sourcing and comparing travel transport options based on user-defined constraints.
Expense & Claims Ops
Automate the extraction, validation, and organization of expense receipts into submission-ready reimbursement packages.
ClawCoach Food
A sophisticated nutrition tracking skill for Openclaw Skills that leverages Claude Vision to analyze food photos and log macros automatically.
ClawCoach Setup
A specialized configuration agent that builds your personalized AI health coaching profile and calculates precise nutritional targets locally.
METADATA
Featured*
