# Add to your Claude Code skills
git clone https://github.com/Gao-Ruilin/AutoRunLast scanned: 5/30/2026
{
"issues": [],
"status": "PASSED",
"scannedAt": "2026-05-30T17:04:28.472Z",
"npmAuditRan": true,
"pipAuditRan": false
}AutoRun is an open-source ai agents skill for AI coding assistants such as Claude Code, Codex CLI, and ChatGPT, built by Gao-Ruilin. An AI agent for coding and others. It has 102 GitHub stars.
Yes. AutoRun passed SkillsLLM's automated security scan — a dependency vulnerability audit plus prompt-injection heuristics — with no high-severity issues. You can read the full report in the Security Report section on this page.
Clone the repository with "git clone https://github.com/Gao-Ruilin/AutoRun" and add it to your Claude Code skills directory (see the Installation section above).
AutoRun is primarily written in Python. It is open-source under Gao-Ruilin on GitHub, so you can review or fork the full source.
Yes. SkillsLLM lists many other AI Agents skills you can browse and compare side by side. Open the AI Agents category from the badge at the top of this page, or use the Related Skills and comparison links further down to weigh AutoRun against similar tools.
No comments yet. Be the first to share your thoughts!
A universal AI coding assistant supporting OpenAI and Anthropic compatible APIs.
Requires Python 3.8+.
Windows:
install.bat
macOS / Linux:
./install.sh
This creates a virtual environment, installs dependencies, and sets up the autorun command.
cd AutoRUN_v1
pip install -e .
After installation, the autorun command is available system-wide:
autorun --version
cd AutoRUN_v1
pip install -r requirements.txt
python cli.py
autorun
On first run, AutoRUN will guide you through API configuration interactively:
╔══════════════════════════════════════════╗
║ AutoRUN v1.0 — 首次设置 ║
╚══════════════════════════════════════════╝
选择 API 类型 [1=OpenAI, 2=Anthropic] [默认 openai]: 1
API URL [默认 https://api.openai.com]:
API Key: sk-xxxxxxxxxxxx
模型名称 [默认 gpt-4o]:
✓ 配置已保存！
Config is stored at ~/.autorun/config.json.
Three methods, listed by priority:
autorun
Follow the interactive prompts on first run, or run autorun --setup to reconfigure anytime.
Windows (PowerShell):
$env:AUTORUN_API_TYPE = "openai"
$env:AUTORUN_API_URL = "https://api.openai.com"
$env:AUTORUN_API_KEY = "sk-xxxxxxxxxxxx"
$env:AUTORUN_MODEL = "gpt-4o"
autorun
Windows (CMD):
set AUTORUN_API_TYPE=openai
set AUTORUN_API_URL=https://api.openai.com
set AUTORUN_API_KEY=sk-xxxxxxxxxxxx
set AUTORUN_MODEL=gpt-4o
autorun
macOS / Linux:
export AUTORUN_API_TYPE=openai
export AUTORUN_API_URL=https://api.openai.com
export AUTORUN_API_KEY=sk-xxxxxxxxxxxx
export AUTORUN_MODEL=gpt-4o
autorun
autorun
Inside the REPL:
/api type openai # API type: openai or anthropic
/api url https://api.openai.com # API base URL
/api key sk-xxxxxxxxxxxx # Your API key
/model gpt-4o # Model name
autorun
autorun --web
Default: http://127.0.0.1:8765
# Custom port and host
autorun --web --port 8080 --host 0.0.0.0
echo "Explain this code" | autorun --print
autorun --print "Explain this code"
autorun --setup
| Flag | Description |
|---|---|
--version, -V, -v |
Show version |
--setup |
Re-run API setup wizard |
--web |
Start Web UI server |
--print, -p |
Pipe/print mode (non-interactive) |
--port <N> |
Web UI port (default: 8765) |
--host <HOST> |
Web UI host (default: 127.0.0.1) |
-m <model>, --model <model> |
Override default model |
-d <path>, --dir <path> |
Working directory |
--context <N> |
Context window size (tokens) |
| Command | Description |
|---|---|
/help |
Show help |
/api |
Show current API config |
/api type <openai|anthropic> |
Set API type |
/api url <URL> |
Set API base URL |
/api key <KEY> |
Set API key |
/model <name> |
Set model name |
/context [tokens] |
Show/set context window size |
/status |
Show session status |
/clear |
Clear conversation history |
/compact |
Compact conversation context |
/memory |
Show memory status |
/skill |
List available skills |
/todos |
Show todo list |
/fast |
Toggle fast mode |
/exit |
Exit |
/v1/chat/completions compatible services (OpenAI, Azure OpenAI, DeepSeek, Groq, etc.)/v1/messages compatible services (Claude series)~/.autorun/
├── config.json # API key, URL, type, model, context window
├── history # Input history
├── memory/ # Memory system
└── skills/ # User skills
.env files)AutoRUN includes a built-in OCR tool based on Falcon-OCR, a lightweight 0.3B vision-language model that runs entirely on your local machine.
HF_ENDPOINT=https://hf-mirror.com).The OCR tool is automatically available in the REPL and Web UI. The AI will call it when you ask it to read text from an image:
> 帮我把这张截图里的文字提取出来
Or use with explicit mode:
> 用 layout 模式 OCR 这篇论文的截图
| Environment Variable | Description | Default |
|---|---|---|
HF_ENDPOINT |
HuggingFace mirror (set to https://hf-mirror.com for China) |
https://huggingface.co |
AUTORUN_OCR_LOCAL_DIR |
Local model directory (skip download) | Auto-download |
AUTORUN_OCR_DEVICE |
Inference device (cuda/cpu) |
Auto-detect |
AUTORUN_OCR_DTYPE |
Model dtype (float32/float16/bfloat16) |
float32 |
torch, transformers, torchvision, huggingface_hub, safetensors, tokenizers, PillowAutoRUN_v1/
├── cli.py # Entry point (autorun command)
├── main.py # CLI argument parsing & routing
├── commands.py # Slash command system
├── query.py # Core query loop
├── query_engine.py # High-level conversation orchestrator
├── pyproject.toml # Package metadata & dependencies
├── requirements.txt # Pip dependencies (legacy)
├── install.bat # Windows one-click installer
├── install.sh # macOS/Linux one-click installer
├── api/ # API clients (OpenAI / Anthropic)
├── context/ # Context building (git, environment, etc.)
├── messages/ # Message types & utilities
├── prompts/ # System prompts
├── services/ # Compaction, LSP, etc.
├── skills/ # Skill discovery & loading
├── state/ # Session state management
├── tools/ # Tool registry & execution
│ └── ocr_engine/ # Falcon-OCR inference engine (Apache-2.0, from TII)
├── ui/ # CLI / Web UI
│ ├── cli/ # prompt_toolkit / Textual REPL
│ └── web/ # FastAPI Web server + frontend
└── utils/ # Config, tokens, env utils
通用 AI 编程助手，支持 OpenAI 和 Anthropic 兼容 API。
需要 Python 3.8+。
Windows:
install.bat
macOS / Linux:
./install.sh
自动创建虚拟环境、安装依赖，并配置 autorun 命令。
cd AutoRUN_v1
pip install -e .
安装后，autorun 命令在终端中全局可用：
autorun --version
cd AutoRUN_v1
pip install -r requirements.txt
python cli.py
autorun
首次运行会自动引导你配置 API：
╔══════════════════════════════════════════╗
║ AutoRUN v1.0 — 首次设置 ║
╚══════════════════════════════════════════╝
选择 API 类型 [1=OpenAI, 2=Anthropic] [默认 openai]: 1
API URL [默认 https://api.openai.com]:
API Key: sk-xxxxxxxxxxxx
模型名称 [默认 gpt-4o]:
✓ 配置已保存！
配置保存在 ~/.autorun/config.json。
三种配置方式，按优先级排序：
autorun
首次运行自动触发，或通过 autorun --setup 随时重新配置。
Windows (PowerShell):
$env:AUTORUN_API_TYPE = "openai"
$env:AUTORUN_API_URL = "https://api.openai.com"
$env:AUTORUN_API_KEY = "sk-xxxxxxxxxxxx"
$env:AUTORUN_MODEL = "gpt-4o"
autorun
Windows (CMD):
set AUTORUN_API_TYPE=openai
set AUTORUN_API_URL=https://api.openai.com
set AUTORUN_API_KEY=sk-xxxxxxxxxxxx
set AUTORUN_MODEL=gpt-4o
autorun
macOS / Linux:
export AUTORUN_API_TYPE=openai
export AUTORUN_API_URL=https://api.openai.com
export AUTORUN_API_KEY=sk-xxxxxxxxxxxx
export AUTORUN_MODEL=gpt-4o
autorun
autorun
进入 REPL 后输入：
/api type openai # API 类型: openai 或 anthropic
/api url https://api.openai.com # API 基础 URL
/api key sk-xxxxxxxxxxxx # 你的 API 密钥
/model gpt-4o # 模型名称
配置保存在 ~/.autorun/config.json，下次启动自动加载。
autorun
autorun --web
# 自定义端口和地址
autorun --web --port 8080 --host 0.0.0.0
echo "解释一下这段代码" | autorun --print
autorun --print "解释一下这段代码"
autorun --setup
| 选项 | 说明 |
|---|---|
--version, -V, -v |
显示版本号 |
--setup |
重新运行 API 设置向导 |
--web |
启动 Web UI 服务器 |
--print, -p |
管道/打印模式（非交互） |
--port <N> |
Web UI 端口（默认: 8765） |
--host <HOST> |
Web UI 地址（默认: 127.0.0.1） |
-m <model>, --model <model> |
覆盖默认模型 |
-d <path>, --dir <path> |
工作目录 |
--context <N> |
上下文窗口大小（tokens） |
| 命令 | 说明 |
|---|---|