blackbox

Name: blackbox
Availability: InStock
Author: icyzh

Delegate coding, refactoring, and PR review tasks to a multi-model AI agent CLI.

1|Updated Jul 31, 2026

Search Tags:#refactoring #multi-agent #terminal #ai-development #coding-agent #blackbox

Ready to run

npx skills add https://github.com/icyzh/hermes-web --skill blackbox-icyzh

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: blackbox
Source: https://github.com/icyzh/hermes-web/tree/main/optional-skills/autonomous-ai-agents/blackbox
Command: npx skills add https://github.com/icyzh/hermes-web --skill blackbox-icyzh

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires @blackbox_ai/blackbox-cli.

What problem does it solve?

This skill solves the challenge of managing complex coding tasks by delegating them to a multi-model AI agent that evaluates and selects the best implementation from various LLMs.

Core Features & Use Cases

Multi-Model Judging: Dispatches coding tasks to multiple models like Claude, Gemini, and Codex, using a judge to select the optimal output.
Interactive & Background Execution: Supports both real-time interactive terminal sessions and long-running background tasks with progress monitoring.
Checkpointing & Resume: Allows developers to save progress on complex refactoring or feature implementation tasks and resume them later.

Quick Start

Use the blackbox skill to add JWT authentication to the current project by running the blackbox command with the appropriate prompt and pty enabled.

Frequently Asked Questions about blackbox

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I automate complex coding and refactoring tasks in the terminal?▼

You can automate complex coding and refactoring tasks in the terminal by delegating them to a multi-model AI agent CLI. This agent evaluates and selects the optimal implementation from multiple LLMs to handle your software development automatically.

How does multi-model judging work for AI coding agents?▼

Multi-model judging works by dispatching your coding prompt to multiple LLMs like Claude, Gemini, and Codex. A judge mechanism then evaluates the generated outputs to select the best implementation for your project.

Do I need a Node.js environment to use the blackbox CLI for PR review?▼

Yes, using the blackbox CLI for PR review requires Node.js 20 or higher and a configured Blackbox AI API key. These prerequisites allow the terminal environment to interface with multiple LLM providers.

What is the best way to manage long-running background coding tasks?▼

The best way to manage long-running background coding tasks is using an AI agent CLI that supports background execution. It provides progress monitoring and checkpoint-based workflow management to save and resume work later.

Can I resume an interactive terminal coding session after stopping it?▼

Yes, you can resume an interactive terminal coding session through checkpointing. The multi-model AI agent allows you to save progress on complex feature implementation or refactoring tasks and resume them later.