DD

devhive1/devhive-cli

Developer tools
18 stars 品質 85 トレンド 85

is a terminal-based for intelligent and autonomous workflows. It allows developers to seamlessly interact with multiple AI agents (Architects, Coders, Debuggers, Orchestrators) inside their terminal...

概要

is a terminal-based for intelligent and autonomous workflows. It allows developers to seamlessly interact with multiple AI agents (Architects, Coders, Debuggers, Orchestrators) inside their terminal...

README

⬡ DevHive CLI

DevHive CLI is a terminal-based Multi-Agent AI Engineering Platform for intelligent and autonomous workflows. It allows developers to seamlessly interact with multiple AI agents (Architects, Coders, Debuggers, Orchestrators) inside their terminal leveraging local and cloud models (via Ollama or supported APIs) to brainstorm, debug, and write code securely.


⚡ What’s New (Latest Updates)

  • Advanced Recursive Agent Loader: DevHive can now automatically crawl and load over 100+ nested agents and skill folders (supporting structures like agents/skill-name/SKILL.md).
  • Heavy LLM Optimizations: Injected CRITICAL RULES FOR AI to strictly enforce tool calls and autonomous execution. This ensures massive reasoning models (like minimax, qwen2.5-coder, and 800b parameters) execute code rapidly without hallucinating, freezing, or stopping mid-task.
  • Smart edit_file Tool: Replaced legacy heavy file-rewriting with precise line-level string replacement, significantly saving context window tokens and drastically speeding up modifications.
  • Import Scripts: Bundled import-agents.js to easily migrate your third-party isolated skills or agents directly into the DevHive core.
  • UI/UX Polish: Upgraded the CLI visual display with stunning, 53-character beautifully aligned Gold ASCII layouts for an immersive dashboard feel.

📸 Screenshots


🔥 Features

  • Multi-Agent Architecture: Comes equipped to handle a massive matrix of custom AI agents. orchestrator, frontend, backend, security, and heavily tailored workflows.
  • Dynamic Command Palette: A robust UI featuring autocomplete slash commands (e.g. /agent, /model, /skills) to easily navigate the tool without leaving your keyboard.
  • MCP Integration (Model Context Protocol): Store and index knowledge, recommend tools, and interact seamlessly with advanced servers.
  • Provider Agnostic: Works perfectly with local Ollama models or any other AI backend connected to the interface.
  • Smart Terminal Environment: Agent context automatically captures your working directory, date, and local environment references.

🚀 Getting Started

Prerequisites

  • Node.js (v16.14 or higher is recommended)
  • Git
  • Ollama (if you plan to use local AI models securely)

Installation

  1. Clone the repository:

    git clone https://github.com/El3tar-cmd/DevHive-Cli.git
    cd DevHive-Cli
    
  2. Install dependencies:

    npm install
    
  3. Migrate/Bundle your Custom Agents (Optional):

    node import-agents.js
    
  4. Build the CLI:

    npm run build
    
  5. Link the CLI globally (Optional but recommended):

    npm link
    

🛠️ Usage

If you used npm link, you can simply run DevHive anywhere using:

hive

(Alternatively, run it via npm run start or node dist/index.js)

Interactive Commands

DevHive CLI comes with a rich set of built-in commands. You can access the command palette securely at any prompt by typing /.

Command Description
/model list View all available models
/model set Switch the active AI model (e.g. minimax-m2.7)
/agent Let the system auto-assign the best specialized agent
/agent Explicitly switch to an agent (e.g. /agent coder)
/skills Manage, inject, and load special system skills
/mcp Interact with the Model Context Protocol engine
/session Manage your active context and chat histories
/clear Clear the current screen and terminal buffer
/exit Exit DevHive

🏗️ Project Architecture

DevHive consists of several key modules:

  • src/ui/: Advanced terminal handling, colored output elements, display widgets, and interactive command palettes.
  • src/agents/: Core definitions, smart caching loaders, and runtime runners handling critical multi-agent tool loops.
  • src/tools/: Powerful operational tools (Shell execution, Fetch, Directory Mgmt, Fast File Editor).
  • src/providers/: AI model connections (Ollama streaming proxy, payload parsers).
  • src/mcp/: Handling interactions with MCP clients and tool calling protocols.

👤 Author

Abdalrahman Mahmoud

📄 License

This project is licensed under the MIT License - feel free to customize and expand DevHive.

View this README on GitHub

推奨ツール

別のキーワードを試すか、フィルタを外してください。

インストール

npx skillfish add devhive1/devhive-cli