agyforclaude
MCP server enabling delegation of tasks to the Antigravity (Gemini) CLI from any AI client supporting MCP.
README
agyforclaude
Universal MCP server + Claude skill for delegating tasks to the Antigravity (Gemini) CLI.
agyforclaude installs an MCP server and a Claude skill into ~/.claude/, making it work with any AI client that supports Model Context Protocol — Claude Code, OpenCode, Cursor, Codex, and more.
Prerequisites
- Node.js >= 18
- Antigravity CLI (
agy) — install it first
Install
npm install -g agyforclaude@latest
The postinstall hook automatically:
- Checks if
agyis installed - Appends
agyforclaudeto~/.claude/mcp.json - Copies the
agyskill to~/.claude/skills/agy/SKILL.md
Then restart your AI agent to pick up the new MCP server.
Usage
Your AI agent now has an agy_delegation tool:
| Param | Type | Default | Description |
|---|---|---|---|
prompt |
string |
(required) | The task to delegate to Gemini |
tier |
flash / flash-lo / pro |
flash |
Model tier |
dir |
string[] |
[] |
Workspace directories to include |
yolo |
boolean |
false |
Auto-approve agy tool permissions |
sandbox |
boolean |
false |
Run agy in sandbox mode |
Example prompts
- "Use agy to scaffold a Next.js app with TypeScript and Tailwind"
- "Delegate to Gemini: generate unit tests for
src/utils.ts" - "Use agy (pro tier) to review this PR for security issues"
The skill
The installed ~/.claude/skills/agy/SKILL.md reminds your agent:
- When to delegate (user asks, bulk work, cheap work)
- Always verify agy's output
- Never auto-merge
CLI Commands
After install, you also get agyforclaude on your PATH:
agyforclaude setup # Re-run setup (e.g. after moving node_modules)
agyforclaude uninstall # Remove from ~/.claude/ (also runs on npm uninstall)
agyforclaude status # Check if everything is properly installed
How it works
- MCP Server (
dist/mcp-server.js): Receives tool calls via stdio, spawnsagy --printwith the user's prompt, returns stdout. - Skill (
~/.claude/skills/agy/SKILL.md): Tells the AI when and how to use theagy_delegationtool. - CLI (
bin/agyforclaude): Manages installation/uninstallation of the above.
Dependencies
@modelcontextprotocol/sdk— MCP server SDKagy(external) — The Antigravity CLI (must be installed separately)
License
MIT
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。