跳到正文
TakoAPI
返回技能列表

agent-zero

由 dowingard 提供

Browser & Automation
2.3k下载量由 clawskills.sh 提供
4星标
0浏览量
0 个赞登录后即可为这个技能点赞
—暂无评分
你的评分

简介

Delegate complex coding, research, or autonomous tasks.

描述

该技能的功能

Bridges Clawdbot and Agent Zero for bidirectional task delegation. Lets you send long-running coding, research, or infrastructure tasks to Agent Zero's autonomous execution environment and receive progress updates back. Supports file attachments, conversation history, and task breakdown with tracked checklists.

Agent Zero runs in a persistent Docker container with self-correction loops, making it better suited for tasks that require many sequential tool calls or iterative builds than a single-shot LLM response.

何时使用

  • Delegating a multi-step API build to Agent Zero
  • Attaching a file for autonomous code review
  • Running long infrastructure tasks inside a persistent Docker environment
  • Breaking a large project into tracked subtasks
  • Asking Agent Zero to report progress back mid-task

示例工作流

输入

User asks: Build a REST API with JWT authentication

智能体
  1. 1

    Skill sends the task to Agent Zero via a0_client.js

  2. 2

    Agent Zero begins iterative coding inside its Docker environment

  3. 3

    Agent Zero reports progress back to Clawdbot via clawdbot_client.js

  4. 4

    User receives incremental status updates

  5. 5

    Final output and files are returned to the Clawdbot session

输出

Completed REST API codebase with JWT auth, delivered with a progress summary

前置要求

  • A0_API_KEY from Agent Zero settings or calculated from runtime ID
  • CLAWDBOT_API_TOKEN for the Clawdbot Gateway
  • Agent Zero instance running (Docker recommended on port 50001)
  • Clawdbot Gateway configured with HTTP chatCompletions endpoint enabled
  • LLM API key configured in Agent Zero (e.g. OpenAI or Gemini key)

安装

clawhub install dowingard/agent-zero-bridge

运行 CLI 命令,或把这段提示词粘贴到 Claude Code、Cursor、Codex 或你在用的任意编程智能体里。

链接