Blackbox CLI for Openclaw

A sophisticated multi-model coding agent that executes tasks across several LLMs and utilizes an internal judge to select the highest quality code implementation.

agungprabowo123
v0.1.0
Mar 13, 2026
0
912
0

Install & Download

1. ClawHub CLI

The fastest way to install a skill directly from the registry.

npx clawhub@latest install blackbox-2

2. Manual Installation

Copy the skill folder to one of these locations

Global
~/.openclaw/skills/
Workspace
<project>/skills/

Priority: Workspace > Local > Bundled

3. Prompt Installation

Copy this prompt to OpenClaw to install it automatically.

Help me install blackbox-2 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 Blackbox CLI?

Blackbox CLI is a powerful terminal-based coding assistant designed to delegate complex development tasks to a variety of high-performance AI models. By integrating with Openclaw Skills, developers can leverage a multi-agent workflow where tasks are dispatched to models like Claude, Gemini, and Blackbox Pro. The core value of this tool lies in its unique judge system, which evaluates multiple generated outputs and picks the best implementation for the specific coding problem at hand.

Built for high-stakes software engineering, the CLI supports interactive terminal sessions and non-interactive one-shot commands. It is particularly effective within the Openclaw Skills ecosystem for users who require more than just a single-model response, offering a robust, open-source framework for code generation, refactoring, and technical analysis.

Blackbox CLI Use Cases

  • Automating the creation of complex API endpoints and database schemas with Openclaw Skills.
  • Refactoring legacy codebases by running logic through multiple LLM providers for optimal results.
  • Conducting deep-dive security audits and PR reviews in temporary directories.
  • Managing long-running architectural changes in background mode with real-time monitoring.
  • Using multimodal vision support to convert design mockups directly into functional frontend code.

How Blackbox CLI Works

  1. The user triggers the tool via a terminal command within the Openclaw Skills environment.
  2. The Blackbox CLI captures the intent from a natural language prompt and identifies the project context.
  3. The task is dispatched to multiple AI models simultaneously if multi-model mode is enabled.
  4. An internal 'Chairman' or judge algorithm analyzes the different model responses.
  5. The optimal code solution is selected and applied to the local filesystem or displayed to the user.
  6. Checkpoints are automatically generated, allowing the user to resume or iterate on the task at a later time.

Blackbox CLI Setup

To get started, ensure you have Node.js 20+ installed. Install the CLI using npm:

npm install -g @blackboxai/cli

Alternatively, you can install from the source repository. Once installed, obtain your API key from the Blackbox AI dashboard and run the configuration command:

blackbox configure

When utilizing this tool as part of your Openclaw Skills setup, always ensure the pty=true flag is enabled in your terminal calls to maintain interactive session support.

Blackbox CLI Data Schema & Taxonomy

Component Description
.blackboxcli/settings.json Local configuration file managing token limits and preferred AI providers.
Checkpoint Tags Unique identifiers (e.g., task-abc123) used to store and resume the state of a coding session.
Session Tokens Usage statistics and history tracked to optimize token consumption and costs.
Workdir Context The local directory path used to scope the AI agent's file system access and knowledge.

Blackbox CLI Advanced Features

  • Multi-model judging workflow that consumes results from several LLMs for maximum reliability.
  • Checkpoint and resume capabilities to handle complex, multi-stage coding projects without losing progress.
  • Background execution mode that allows developers to monitor long tasks via logs while performing other Openclaw Skills actions.
  • YOLO mode for power users who want the AI to auto-approve all file changes and model switches.
  • Vision-integrated analysis that switches to multimodal models when images or UI screenshots are detected in the input.
  • Session compression commands to manage token usage during long-form technical conversations.

SKILL.md


Loading

Related Openclaw Skills

METADATA

Github Stars: 0
forks: 0

Featured*