自动化AI工作流 是 AI Skill Hub 本期精选Agent工作流之一。综合评分 8.0 分,整体质量较高。我们强烈推荐将其纳入你的 AI 工具库,帮助提升工作效率。
自动化AI工作流 是一套完整的 AI Agent 自动化工作流方案。通过可视化的节点编排,将复杂的多步骤任务拆解为清晰的自动化流程,实现全程无人值守的智能处理。支持与数百种外部服务和 API 无缝集成,适合构建数据处理管线、业务自动化和 AI 辅助决策系统。
自动化AI工作流 是一套完整的 AI Agent 自动化工作流方案。通过可视化的节点编排,将复杂的多步骤任务拆解为清晰的自动化流程,实现全程无人值守的智能处理。支持与数百种外部服务和 API 无缝集成,适合构建数据处理管线、业务自动化和 AI 辅助决策系统。
# 方式一:pip 安装(推荐)
pip install autobots
# 方式二:虚拟环境安装(推荐生产环境)
python -m venv .venv
source .venv/bin/activate # Windows: .venv\Scripts\activate
pip install autobots
# 方式三:从源码安装(获取最新功能)
git clone https://github.com/DanielDeshmukh/autobots
cd autobots
pip install -e .
# 验证安装
python -c "import autobots; print('安装成功')"
# 命令行使用
autobots --help
# 基本用法
autobots input_file -o output_file
# Python 代码中调用
import autobots
# 示例
result = autobots.process("input")
print(result)
# autobots 配置文件示例(config.yml) app: name: "autobots" debug: false log_level: "INFO" # 运行时指定配置文件 autobots --config config.yml # 或通过环境变量配置 export AUTOBOTS_API_KEY="your-key" export AUTOBOTS_OUTPUT_DIR="./output"
<p align="center"> <a href="https://github.com/DanielDeshmukh/autobots"> <img src="https://img.shields.io/badge/version-0.2.0-blue" alt="Version"> </a> <a href="https://python.org"> <img src="https://img.shields.io/badge/python-3.11+-brightgreen" alt="Python"> </a> <a href="https://pypi.org/project/autobot-swarm/"> <img src="https://img.shields.io/pypi/v/autobot-swarm" alt="PyPI"> </a> <a href="https://github.com/DanielDeshmukh/autobots/blob/main/LICENSE"> <img src="https://img.shields.io/badge/license-MIT-blue" alt="License"> </a> </p>
Autobots is a hierarchical multi-cluster coding swarm CLI that orchestrates multiple AI models to plan, implement, validate, and repair code against target repositories. It features a full tool system (Read, Write, Edit, Glob, Grep), interactive REPL mode, granular permissions, context management, hooks, MCP integration, and code review — all built on NVIDIA NIM models.
---
pip install autobot-swarm
git clone https://github.com/DanielDeshmukh/autobots.git
cd autobots
python -m pip install -e . --no-build-isolation
---
```powershell
```powershell
echo NVIDIA_API_KEY=your_key_here > .env
$env:NVIDIA_API_KEY = "your_key_here" ```
Create .autobots.toml or autobots.toml in your project root or $HOME:
```toml [autobots] model_selection_profile = "balanced" # balanced, speed, or quality parallel_planning = false disable_live_catalog = false safety_branch = "autobots-safety" default_mode = "supervised" # supervised, milestone, autonomous milestone_threshold = 3 max_verification_attempts = 3 temperature = 0.2 max_tokens = 4096
Create ~/.autobots/settings.json (global) or .autobots/settings.json (project):
{
"permissions": {
"default": "ask",
"rules": [
{"tool_pattern": "Read", "permission": "allow"},
{"tool_pattern": "Glob", "permission": "allow"},
{"tool_pattern": "Grep", "permission": "allow"},
{"tool_pattern": "Write", "permission": "ask"},
{"tool_pattern": "Edit", "permission": "ask"},
{"tool_pattern": "Bash", "permission": "ask"}
]
}
}
---
Global (~/.autobots/settings.json):
{
"permissions": {
"default": "ask",
"rules": [
{"tool_pattern": "Read", "permission": "allow"},
{"tool_pattern": "Write", "permission": "ask"}
]
}
}
Project (.autobots/settings.json):
{
"permissions": {
"rules": [
{"tool_pattern": "Bash(npm test)", "permission": "allow"},
{"tool_pattern": "Bash(rm *)", "permission": "deny"}
]
}
}
Environment Variables:
$env:AUTOBOTS_ALLOWED_TOOLS = "Read,Glob,Grep"
$env:AUTOBOTS_DENIED_TOOLS = "Bash(rm *)"
---
| Variable | Description | Default |
|---|---|---|
NVIDIA_API_KEY | NVIDIA API key for model access | Required |
AUTOBOTS_MODEL_SELECTION_PROFILE | Model selection: balanced/speed/quality | balanced |
AUTOBOTS_ENABLE_PARALLEL_PLANNING | Enable parallel workstreams | false |
AUTOBOTS_DISABLE_LIVE_CATALOG | Use bundled models only | false |
AUTOBOTS_SAFETY_BRANCH | Required git branch name | autobots-safety |
AUTOBOTS_DEFAULT_MODE | Default execution mode | supervised |
AUTOBOTS_MILESTONE_THRESHOLD | Phases per approval | 3 |
AUTOBOTS_MAX_VERIFICATION_ATTEMPTS | Retry limit | 3 |
AUTOBOTS_ALLOWED_TOOLS | Comma-separated allowed tools | (none) |
AUTOBOTS_DENIED_TOOLS | Comma-separated denied tools | (none) |
---
$env:NVIDIA_API_KEY = "your_key_here"
| Command | Description |
|---|---|
autobots init | Check context files in target project |
autobots init --interactive | Interactive setup wizard |
autobots plan | Generate implementation roadmap |
autobots plan --goal "text" | Plan with specific goal |
autobots plan --append | Append to existing plan |
autobots plan --dry-run | Preview without writing |
autobots run [task] | Execute phases with autonomy mode |
autobots run --supervised | Manual approval per phase (default) |
autobots run --milestone | Approval every N phases |
autobots run --autonomous | Fully autonomous execution |
autobots resume | Resume from last checkpoint |
autobots engage | Interactive mode with startup screen |
autobots ask "question" | One-shot question answering |
autobots steer "instruction" | Add mid-task steering instructions |
autobots status | Rich status with progress bars |
autobots explain <id> | Show audit trail for a phase/task |
autobots stats | Usage statistics and costs |
autobots undo | Rollback to previous snapshot |
autobots snapshots | List available snapshots |
autobots diff | Compare workspace to snapshot |
autobots logs | View audit trail |
autobots doctor | Preflight health checks |
autobots catalog | Browse NVIDIA model registry |
autobots config validate | Validate TOML configuration |
autobots completions | Generate shell completions |
autobots marketplace | Browse skill packs |
autobots dashboard | Launch web dashboard |
autobots validate-models | Test NVIDIA API connectivity |
autobots publish | Build and publish to PyPI |
---
src/, app/, lib/, tests/, docs/, scripts/---
Connect to Model Context Protocol servers:
from autobots.mcp import MCPClient
client = MCPClient(command="npx", args=["-y", "@modelcontextprotocol/server-filesystem"])
client.connect()
tools = client.list_tools()
result = client.call_tool("read_file", {"path": "/etc/hosts"})
---
| Error | Solution |
|---|---|
NVIDIA_API_KEY is missing | Set NVIDIA_API_KEY in .env or environment variable |
Missing: roadmap.md | Run autobots plan to generate context files |
Switch to autobots-safety branch | git checkout -b autobots-safety |
Command not in safety whitelist | Autobots blocks dangerous commands. Use safe alternatives |
No checkpoint found | Start fresh with autobots run |
old_string not found | Provide exact surrounding lines for unique match |
old_string found N times | Add more context or use replace_all=true |
---
Autobots是一个创新性的AI工作流项目,具有较高的潜力
AI Skill Hub 为第三方内容聚合平台,本页面信息基于公开数据整理,不对工具功能和质量作任何法律背书。
建议在沙箱或测试环境中充分验证后,再部署至生产环境,并做好必要的安全评估。
✅ MIT 协议 — 最宽松的开源协议之一,可自由商用、修改、分发,仅需保留版权声明。
经综合评估,自动化AI工作流 在Agent工作流赛道中表现稳健,质量优秀。如果你已有明确的使用需求,可以直接上手体验;如果还在评估阶段,建议对比同类工具后再做决策。
| 原始名称 | autobots |
| Topics | AI工作流去中心化代理群 |
| GitHub | https://github.com/DanielDeshmukh/autobots |
| License | MIT |
| 语言 | Python |
收录时间:2026-06-26 · 更新时间:2026-06-26 · License:MIT · AI Skill Hub 不对第三方内容的准确性作法律背书。
选择 Agent 类型,复制安装指令后粘贴到对应客户端