Streamline AI-driven actions with a lightweight REST and CLI-based approval system for Openclaw Skills.
The fastest way to install a skill directly from the registry.
npx clawhub@latest install approval-queue
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
Copy this prompt to OpenClaw to install it automatically.
Help me install approval-queue using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Get the raw skill files in a ZIP archive.
The Approval Queue skill provides a robust human-in-the-loop mechanism to manage autonomous actions generated by agents. Whether it is social media posts, software deployments, or content publishing, this skill ensures that nothing goes live without explicit human authorization. Designed as a lightweight service, it fits perfectly into the ecosystem of Openclaw Skills, offering both a RESTful API and a CLI interface to integrate seamlessly with existing developer workflows.
By implementing this skill, developers can add a layer of safety and governance to their AI agents. The system utilizes a SQLite backend to track pending tasks, allowing users to review, approve, or reject actions with optional reasoning. This makes it an essential tool for maintaining quality control and security in automated environments powered by Openclaw Skills.
To get started, navigate to the skill directory and initialize the environment:
cd {skill_dir}
npm install && npm run build
Start the API server to begin managing your Openclaw Skills actions:
node dist/server.js --port 3010
Use the CLI to add a test item to your queue:
node dist/cli.js add --type sns_post --payload '{"text":"Hello world","platform":"twitter"}'
The system manages data using a structured JSON schema within a SQLite database to ensure compatibility with other Openclaw Skills. Each item in the queue follows this format:
| Field | Type | Description |
|---|---|---|
| id | UUID | Unique identifier for the action item. |
| type | String | The category of action (e.g., sns_post, deploy). |
| status | String | Current state: pending, approved, or rejected. |
| payload | Object | The actual data or command to be executed. |
| created_at | ISO8601 | Timestamp of item creation. |
| reviewer_note | String | Optional feedback or reason for rejection. |
Configuration is handled via environment variables like PORT, DB_PATH, and WEBHOOK_URL.
Loading
An automated log parsing utility that identifies error patterns, calculates frequency distributions, and provides actionable summaries from system and application logs.

A comprehensive tracking and analysis tool designed to monitor performance, engagement, and growth across multiple social media platforms.

A comprehensive automation skill for centralizing affiliate links, tracking performance, and managing multi-platform link insertion with built-in compliance.

A centralized management hub for storing and organizing social media handles, marketplace profiles, and secure credentials for AI agents.

An automated cryptocurrency trading tool designed to detect and execute price arbitrage opportunities across multiple exchanges.

An intelligent error-correction engine that automatically repairs linting, type mismatches, and syntax errors across multiple programming languages.








































