Token导航 LogoToken导航

Agent Skill · 研究检索

antigravity-manager(反重力经理)

antigravity-manager 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

antigravity-manager 是研究检索类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为接触敏感数据;许可证信息还不明确,第三方扫描有未通过项,按安装量在研究检索类的 73,675 个 Skill 里排第 5,155。

总安装 27,456周安装 1,073GitHub Stars 39下载 8,888核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
unknown

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:antigravity-manager(反重力经理)
来源仓库:https://github.com/aradotso/trending-skills
仓库路径:skills/antigravity-manager
安装命令:
npx skills add https://github.com/aradotso/trending-skills --skill antigravity-manager
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/aradotso/trending-skills --skill antigravity-manager

这个 Skill 是做什么的

研究检索类 Skill 主要解决「找资料、核对事实、把来源整理清楚」的问题,通常会让助手先检索再总结。

用于查找、检索和筛选相关信息,提升内容定位效率。适合在 Codex、Claude、Cursor、Gemini CLI 中根据关键词或任务场景快速获取候选结果。主要功能是将 Google 与 Anthropic 的会话令牌转换为标准 API 端点,支持多账户轮换。

安装命令为 npx skills add https://github.com/aradotso/trending-skills --skill antigravity-manager。使用前需确认权限范围及是否涉及敏感凭证管理与网络请求。

来源仓库内路径:skills/antigravity-manager

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/aradotso/trending-skills --skill antigravity-m…;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它需要接触敏感数据,先在测试目录或测试账号里跑,确认没问题再放到正式项目。

常见问题

6 条

antigravity-manager 怎么安装?

在终端执行 npx skills add https://github.com/aradotso/trending-skills --skill antigravity-manager。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

antigravity-manager 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

antigravity-manager 需要什么权限?

权限级别是接触敏感数据。该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

antigravity-manager 是免费的吗?

公开资料没有标注明确许可证,收费和商用限制都无法确认。正式使用前请到来源仓库核对授权说明。

antigravity-manager 用的人多吗?

公开数据是累计安装 27,456 次,近一周安装 1,073 次,下载 8,888 次,GitHub 39 星。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

antigravity-manager 安全吗?

第三方扫描有未通过项,风险较高。除非你确认过源码,否则不建议直接安装到常用环境。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。

先想清楚的地方

  • 需要接触敏感数据,装到常用环境前要先评估影响面。
  • 许可证没有明确标注,能不能商用、能不能再分发都不确定。
  • 第三方扫描存在提醒项或未通过项,需要人工逐条确认。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理研究检索类任务、又不想每次都手写提示词的人。

不太适合

  • 环境里有大量密钥、凭证,不方便给第三方 Skill 读取的人。
  • 需要明确授权才能用到公司项目里的场景。
  • 需要厂商级技术支持和 SLA 的生产系统。

研究检索类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
find-skills

寻找技能 · Codex、Claude

31,199,81316,486操作浏览器MIT有提醒
self-improving-agent

自我改进 Agent · OpenClaw

16,464,0023,396接触敏感数据MIT-0通过
soultrace

灵魂追踪 · Codex、Claude

7,588,94412需要联网未明确有提醒
self-improving

自我完善 · OpenClaw

6,829,2841,046需要联网MIT-0通过
microsoft-foundry

微软代工厂 · Codex、Claude

6,614,144764调用外部服务未明确有提醒
antigravity-manager

反重力经理 · 本页

27,45639接触敏感数据未明确未通过

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

Antigravity Manager

Skill by ara.so — Daily 2026 Skills collection.

Antigravity Manager is a professional AI account manager and proxy gateway. It takes Google (Gemini) and Anthropic (Claude) web session tokens and exposes them as standard API endpoints (OpenAI-compatible, Anthropic-native, and Gemini-native formats) with intelligent multi-account rotation, quota tracking, and automatic failover.

Key capabilities:

  • Multi-account management with real-time quota dashboards
  • Protocol conversion: web sessions → OpenAI / Anthropic / Gemini API
  • Auto-rotation on 429/401 errors (millisecond failover)
  • Model routing and remapping
  • Desktop app (Tauri v2 + React + Rust) or headless Docker/server mode

Installation

Option A: One-line script (Linux/macOS)

curl -fsSL https://raw.githubusercontent.com/lbjlaq/Antigravity-Manager/v4.1.30/install.sh | bash

Install a specific version:

curl -fsSL https://raw.githubusercontent.com/lbjlaq/Antigravity-Manager/v4.1.30/install.sh | bash -s -- --version 4.1.30

Dry run (preview without installing):

curl -fsSL https://raw.githubusercontent.com/lbjlaq/Antigravity-Manager/v4.1.30/install.sh | bash -s -- --dry-run

Option B: Windows (PowerShell)

irm https://raw.githubusercontent.com/lbjlaq/Antigravity-Manager/main/install.ps1 | iex

Option C: Homebrew (macOS / Linuxbrew)

brew tap lbjlaq/antigravity-manager https://github.com/lbjlaq/Antigravity-Manager
brew install --cask antigravity-tools

Option D: Docker (recommended for servers/NAS)

docker run -d --name antigravity-manager \
  -p 8045:8045 \
  -e API_KEY=$ANTIGRAVITY_API_KEY \
  -e WEB_PASSWORD=$ANTIGRAVITY_WEB_PASSWORD \
  -e ABV_MAX_BODY_SIZE=104857600 \
  -v ~/.antigravity_tools:/root/.antigravity_tools \
  lbjlaq/antigravity-manager:latest

Docker Compose:

# docker-compose.yml
version: "3.8"
services:
  antigravity:
    image: lbjlaq/antigravity-manager:latest
    container_name: antigravity-manager
    restart: unless-stopped
    ports:
      - "8045:8045"
    environment:
      - API_KEY=${ANTIGRAVITY_API_KEY}
      - WEB_PASSWORD=${ANTIGRAVITY_WEB_PASSWORD}
      - ABV_MAX_BODY_SIZE=104857600
    volumes:
      - ~/.antigravity_tools:/root/.antigravity_tools
docker compose up -d
docker logs antigravity-manager          # view logs / recover forgotten keys

Option E: Manual download

Download from GitHub Releases:

  • macOS: .dmg (Apple Silicon + Intel)
  • Windows: .msi or portable .zip
  • Linux: .deb, .rpm, or .AppImage

Authentication / Security Model

Antigravity uses two separate credentials:

CredentialEnv varConfig keyPurpose
API KeyAPI_KEYapi_keyAuthenticates AI API calls (Authorization: Bearer...)
Web PasswordWEB_PASSWORDadmin_passwordLogs into the management web UI

Scenario A — only API_KEY set:

  • Web UI login: use API_KEY
  • API calls: use API_KEY

Scenario B — both set (recommended):

  • Web UI login: WEB_PASSWORD only (API Key rejected for login)
  • API calls: API_KEY only

Recover credentials:

docker logs antigravity-manager
# or
grep -E '"api_key"|"admin_password"' ~/.antigravity_tools/gui_config.json

API Endpoints

The proxy server runs on port 8045 by default.

OpenAI-compatible (works with any OpenAI SDK)

POST http://localhost:8045/v1/chat/completions
Authorization: Bearer $ANTIGRAVITY_API_KEY

Anthropic-native (Claude Code, etc.)

POST http://localhost:8045/v1/messages
Authorization: Bearer $ANTIGRAVITY_API_KEY

Gemini-native

POST http://localhost:8045/v1/models/{model}:generateContent
Authorization: Bearer $ANTIGRAVITY_API_KEY

Code Examples

Python — OpenAI SDK (Gemini via Antigravity)

import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["ANTIGRAVITY_API_KEY"],
    base_url="http://localhost:8045/v1",
)

response = client.chat.completions.create(
    model="gemini-2.5-pro",
    messages=[
        {"role": "user", "content": "Explain quantum entanglement in simple terms."}
    ],
    stream=True,
)

for chunk in response:
    if chunk.choices[0].delta.content:
        print(chunk.choices[0].delta.content, end="", flush=True)

Python — Anthropic SDK (Claude via Antigravity)

import os
import anthropic

client = anthropic.Anthropic(
    api_key=os.environ["ANTIGRAVITY_API_KEY"],
    base_url="http://localhost:8045",
)

message = client.messages.create(
    model="claude-sonnet-4-5",
    max_tokens=1024,
    messages=[
        {"role": "user", "content": "Write a haiku about distributed systems."}
    ],
)
print(message.content[0].text)

Python — Streaming with Anthropic

import os
import anthropic

client = anthropic.Anthropic(
    api_key=os.environ["ANTIGRAVITY_API_KEY"],
    base_url="http://localhost:8045",
)

with client.messages.stream(
    model="claude-opus-4-5",
    max_tokens=2048,
    system="You are a helpful coding assistant.",
    messages=[{"role": "user", "content": "Implement a binary search in Rust."}],
) as stream:
    for text in stream.text_stream:
        print(text, end="", flush=True)

TypeScript / Node — OpenAI SDK

import OpenAI from "openai";

const client = new OpenAI({
  apiKey: process.env.ANTIGRAVITY_API_KEY!,
  baseURL: "http://localhost:8045/v1",
});

async function chat(prompt: string): Promise<string> {
  const response = await client.chat.completions.create({
    model: "gemini-2.5-flash",
    messages: [{ role: "user", content: prompt }],
  });
  return response.choices[0].message.content ?? "";
}

// Image generation via Imagen 3
async function generateImage(prompt: string) {
  const response = await client.images.generate({
    model: "imagen-3.0-generate-002",
    prompt,
    size: "1024x1024",  // maps to Imagen 3 aspect ratio
    n: 1,
  });
  return response.data[0].url;
}

cURL — quick test

# Test OpenAI-compatible endpoint
curl -s http://localhost:8045/v1/chat/completions \
  -H "Authorization: Bearer $ANTIGRAVITY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gemini-2.5-pro",
    "messages": [{"role": "user", "content": "Hello!"}]
  }' | jq '.choices[0].message.content'

# Test Anthropic endpoint
curl -s http://localhost:8045/v1/messages \
  -H "Authorization: Bearer $ANTIGRAVITY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "claude-sonnet-4-5",
    "max_tokens": 256,
    "messages": [{"role": "user", "content": "Hello!"}]
  }' | jq '.content[0].text'

Connecting Claude Code CLI

Claude Code uses the Anthropic protocol. Point it at Antigravity:

# Set environment variables before starting Claude Code
export ANTHROPIC_API_KEY=$ANTIGRAVITY_API_KEY
export ANTHROPIC_BASE_URL=http://localhost:8045

claude  # start Claude Code — it will use Antigravity as the backend

Or create a .env in your project root:

# .env
ANTHROPIC_API_KEY=your-antigravity-api-key
ANTHROPIC_BASE_URL=http://localhost:8045

Connecting to Popular AI Clients

Cherry Studio

  1. Open Cherry Studio → Settings → API Provider
  2. Select OpenAI-compatible
  3. Base URL: http://localhost:8045/v1
  4. API Key: your ANTIGRAVITY_API_KEY

Cursor

In Cursor settings, set:

  • OpenAI Base URL: http://localhost:8045/v1
  • API Key: your ANTIGRAVITY_API_KEY

Continue.dev

// ~/.continue/config.json
{
  "models": [
    {
      "title": "Gemini Pro (Antigravity)",
      "provider": "openai",
      "model": "gemini-2.5-pro",
      "apiKey": "YOUR_ANTIGRAVITY_API_KEY",
      "apiBase": "http://localhost:8045/v1"
    },
    {
      "title": "Claude (Antigravity)",
      "provider": "anthropic",
      "model": "claude-sonnet-4-5",
      "apiKey": "YOUR_ANTIGRAVITY_API_KEY",
      "apiBase": "http://localhost:8045"
    }
  ]
}

Configuration File

Config is stored at ~/.antigravity_tools/gui_config.json:

{
  "api_key": "sk-your-api-key",
  "admin_password": "your-web-ui-password",
  "proxy": {
    "port": 8045,
    "max_body_size": 104857600,
    "admin_password": "your-web-ui-password"
  },
  "model_routes": [
    {
      "pattern": "gpt-4.*",
      "target": "gemini-2.5-pro"
    },
    {
      "pattern": "gpt-3.5.*",
      "target": "gemini-2.5-flash"
    }
  ]
}

Environment Variables (Docker / server mode)

VariableDescriptionDefault
API_KEYAPI key for authenticating proxy requestsrequired
WEB_PASSWORDWeb UI admin passwordfalls back to API_KEY
ABV_MAX_BODY_SIZEMax request body in bytes (for image uploads)104857600 (100MB)

Model Routing

Antigravity's model router lets you remap incoming model names to actual upstream models. This is useful when clients hard-code model names like gpt-4.

Configuration via Web UI (API Proxy → Model Router) or gui_config.json:

{
  "model_routes": [
    {
      "pattern": "gpt-4-turbo",
      "target": "gemini-2.5-pro"
    },
    {
      "pattern": "gpt-4o",
      "target": "gemini-2.5-pro"
    },
    {
      "pattern": "gpt-3\\.5.*",
      "target": "gemini-2.5-flash"
    },
    {
      "pattern": "claude-3-opus.*",
      "target": "claude-opus-4-5"
    }
  ]
}

Tiered routing (automatic): Antigravity prioritizes accounts by type (Ultra → Pro → Free) and quota reset frequency, consuming fast-reset accounts first.

Background task demotion (automatic): Requests detected as background tasks (e.g., Claude CLI title generation) are automatically redirected to Flash-tier models to preserve premium quota.


Account Management

Adding accounts via the UI

  1. Open Antigravity Manager desktop app or web UI (http://localhost:8045)
  2. Navigate to Accounts
  3. Click Add Account — the app generates an OAuth 2.0 authorization URL
  4. Complete authorization in any browser
  5. The app captures the callback automatically; click "I've authorized, continue" if needed

Batch import (JSON)

Export from another tool and import in the Accounts page:

[
  {
    "token": "session-token-1",
    "type": "google",
    "label": "account-work"
  },
  {
    "token": "session-token-2",
    "type": "anthropic",
    "label": "account-personal"
  }
]

Migration from v1

Antigravity auto-detects and migrates v1 database format on first launch — no manual steps needed.


Quota Monitoring

The dashboard shows live quota across all accounts:

  • Gemini Pro remaining calls (average across accounts)
  • Gemini Flash remaining calls
  • Claude remaining calls
  • Imagen 3 image generation quota

Smart Recommendation: The dashboard algorithmically selects the account with the most headroom and offers one-click switching.

403 detection: Accounts banned upstream are automatically flagged and skipped during rotation.


Building from Source (Rust + Tauri v2)

Prerequisites

# Rust
curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh

# Node.js (for frontend)
node --version  # requires v18+

# Tauri v2 CLI
cargo install tauri-cli --version "^2"

# System deps (Linux)
sudo apt install libwebkit2gtk-4.1-dev libayatana-appindicator3-dev librsvg2-dev

Build & run

git clone https://github.com/lbjlaq/Antigravity-Manager.git
cd Antigravity-Manager

npm install          # install frontend deps
cargo tauri dev      # run in development mode
cargo tauri build    # build release binaries → src-tauri/target/release/bundle/

Build headless server only (no GUI)

cd src-tauri
cargo build --release --features headless
./target/release/antigravity-manager --headless --port 8045

Troubleshooting

Port 8045 already in use

# Find and kill the process
lsof -ti:8045 | xargs kill -9
# or change port in gui_config.json:
# "proxy": { "port": 8046 }

Forgotten API key or web password

# Docker
docker logs antigravity-manager | grep -E "api_key|password"

# Local install
cat ~/.antigravity_tools/gui_config.json | grep -E '"api_key"|"admin_password"'

Account shows 403 / banned

The account has been flagged upstream. In the Accounts view it will be marked with a 403 badge and automatically skipped. Add a fresh account and remove/re-authorize the banned one.

Requests failing with 429

Antigravity handles this automatically (rotates to next account within milliseconds). If all accounts are exhausted, add more accounts or wait for quota reset. Check the dashboard for per-account quota status.

macOS Gatekeeper blocking the app

xattr -d com.apple.quarantine /Applications/Antigravity\ Tools.app
# or install with:
brew install --cask antigravity-tools --no-quarantine

Docker container not persisting accounts after restart

Ensure the volume mount is correct:

docker run ... -v ~/.antigravity_tools:/root/.antigravity_tools ...
# Verify data is present:
ls ~/.antigravity_tools/

Claude Code not connecting

# Verify the proxy is running
curl -s http://localhost:8045/v1/models \
  -H "Authorization: Bearer $ANTIGRAVITY_API_KEY"

# Check env vars are exported (not just set)
export ANTHROPIC_BASE_URL=http://localhost:8045
export ANTHROPIC_API_KEY=$ANTIGRAVITY_API_KEY
echo $ANTHROPIC_BASE_URL  # should print the URL

Project Structure (for contributors)

Antigravity-Manager/
├── src/                    # React frontend (TypeScript)
│   ├── components/         # UI components (Dashboard, Accounts, etc.)
│   └── pages/
├── src-tauri/              # Rust backend
│   ├── src/
│   │   ├── main.rs         # Tauri app entry
│   │   ├── proxy/          # Axum HTTP proxy server
│   │   │   ├── router.rs   # Model routing logic
│   │   │   ├── dispatcher.rs # Account rotation
│   │   │   └── mapper.rs   # Protocol conversion
│   │   ├── accounts/       # Account storage & OAuth
│   │   └── quota/          # Quota tracking
│   └── Cargo.toml
├── docker/
│   └── Dockerfile
├── install.sh              # Linux/macOS installer
├── install.ps1             # Windows installer
└── docker-compose.yml

平台分布

Codex

39.75%
按下载量换算3,533

Claude

28.93%
按下载量换算2,571

Cursor

19.69%
按下载量换算1,750

Gemini CLI

9.48%
按下载量换算843

安全审计

Gen Agent Trust Hub

未通过

Socket

可疑

Snyk

未通过

权限和风险

接触敏感数据

该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。来源安全扫描存在提醒或未通过项,安装前请自行复核。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills