Antigravity GLM MCP

Antigravity GLM MCP

A bridge between Gemini and GLM-4.5 models that provides 25 tools for automating complex coding tasks, file management, and system operations. It features a security-hardened architecture with sandboxing, persistent memory, and automatic file backups.

Category
Visit Server

README

<div align="center">

πŸš€ Antigravity GLM MCP

Gemini (Antigravity) ⟷ GLM-4.5 λͺ¨λΈ λΈŒλ¦Ώμ§€ MCP μ„œλ²„

λ³΅μž‘ν•œ μ½”λ”© μž‘μ—…μ„ GLM AIμ—κ²Œ μœ„μž„ν•˜κ³ , 25κ°€μ§€ κ°•λ ₯ν•œ λ„κ΅¬λ‘œ μžλ™ν™”ν•˜μ„Έμš”.

Python MCP Tools License Security

πŸ“– λ¬Έμ„œ Β· ⚑ λΉ λ₯Έ μ‹œμž‘ Β· πŸ› οΈ 도ꡬ Β· πŸ›‘οΈ λ³΄μ•ˆ

</div>


✨ μ™œ 이 ν”„λ‘œμ νŠΈμΈκ°€μš”?

문제점 ν•΄κ²°μ±…
🐳 λ³΅μž‘ν•œ Docker μ„€μ • βœ… Zero-Docker: HTTPS 직접 호좜둜 μ¦‰μ‹œ μ‹œμž‘
πŸ” API ν‚€ 유좜 우렀 βœ… λ³΄μ•ˆ κ°•ν™”: ν™˜κ²½λ³€μˆ˜ 필터링, μƒŒλ“œλ°•μŠ€ 적용
πŸ“ 파일 μ‹€μˆ˜ 볡ꡬ λΆˆκ°€ βœ… μžλ™ λ°±μ—…: μˆ˜μ •/μ‚­μ œ μ „ 버전 관리
🧠 μ„Έμ…˜ κ°„ 정보 손싀 βœ… 영ꡬ λ©”λͺ¨λ¦¬: JSON 기반 μž₯κΈ° κΈ°μ–΅ μ €μž₯μ†Œ
⚠️ μœ„ν—˜ν•œ μ‰˜ λͺ…λ Ή μ‹€ν–‰ βœ… ν™”μ΄νŠΈλ¦¬μŠ€νŠΈ: 승인된 λͺ…λ Ήλ§Œ ν—ˆμš©

πŸ—οΈ μ•„ν‚€ν…μ²˜

flowchart LR
    subgraph "πŸ–₯️ Gemini Desktop"
        A[Antigravity Agent]
    end
    
    subgraph "πŸ”§ MCP Server"
        B[antigravity_glm_mcp]
        C[25 Tools]
        D[Security Layer]
    end
    
    subgraph "☁️ Cloud APIs"
        E[GLM-4.5 API]
        F[Web Search]
    end
    
    subgraph "πŸ’Ύ Local Storage"
        G[Files & Git]
        H[Memory DB]
        I[Backups]
    end

    A <--> |MCP Protocol| B
    B --> C
    C --> D
    D --> E
    D --> F
    D --> G
    D --> H
    D --> I

πŸ“– λ¬Έμ„œ

πŸ“„ λ¬Έμ„œ πŸ” μ„€λͺ…
πŸ›οΈ μ•„ν‚€ν…μ²˜ μ‹œμŠ€ν…œ 섀계, λ³΄μ•ˆ 계측, 톡신 흐름 상세
πŸ“š 도ꡬ 레퍼런슀 25개 도ꡬ νŒŒλΌλ―Έν„° 및 응닡 λͺ…μ„Έ
⚑ ν€΅μŠ€νƒ€νŠΈ 5λΆ„ μ„€μΉ˜ κ°€μ΄λ“œ 및 첫 μ‚¬μš© 예제

⚑ λΉ λ₯Έ μ‹œμž‘ (5λΆ„)

πŸ“‹ 사전 μš”κ΅¬μ‚¬ν•­

  • Python 3.11+ (κ°€μƒν™˜κ²½ ꢌμž₯)
  • Zhipu AI API ν‚€ λ˜λŠ” ν˜Έν™˜ μ—”λ“œν¬μΈνŠΈ

πŸš€ μ„€μΉ˜ 방법

# 1. μ €μž₯μ†Œ 클둠
git clone https://github.com/coreline-ai/antigravity_glm_mcp.git
cd antigravity_glm_mcp

# 2. κ°€μƒν™˜κ²½ 생성 및 ν™œμ„±ν™”
python3.11 -m venv .venv
source .venv/bin/activate  # Windows: .venv\Scripts\activate

# 3. μžλ™ μ„€μΉ˜ (ꢌμž₯)
python scripts/install.py

<details> <summary><b>πŸ“ μˆ˜λ™ μ„€μΉ˜ (λŒ€μ•ˆ)</b></summary>

# μ˜μ‘΄μ„± μ„€μΉ˜
pip install -r requirements.txt

# MCP μ„€μ • νŒŒμΌμ— μΆ”κ°€ (~/.gemini/settings.json λ“±)
{
  "mcpServers": {
    "antigravity_glm_mcp": {
      "command": "/path/to/.venv/bin/python",
      "args": ["/path/to/antigravity_glm_mcp/src/server.py"],
      "env": {
        "PROJECT_ROOT": "/your/workspace",
        "ZHIPU_API_KEY": "your-api-key",
        "GLM_MODEL": "GLM-4.5",
        "GLM_BASE_URL": "https://api.z.ai/api/coding/paas/v4",
        "PYTHONPATH": "/path/to/antigravity_glm_mcp"
      }
    }
  }
}

</details>


πŸ› οΈ 전체 도ꡬ λͺ©λ‘ (25개)

🧠 μ§€λŠ₯ μœ„μž„ (Intelligence Delegation)

도ꡬ μ„€λͺ… μ£Όμš” νŒŒλΌλ―Έν„°
glm_cmd GLM에 λ³΅μž‘ν•œ μž‘μ—… μœ„μž„ task_description, context
glm_bypass μ›μ‹œ ν”„λ‘¬ν”„νŠΈ 직접 전솑 prompt
glm_image_analyze 이미지 뢄석 (Vision) image_path, prompt

πŸ“ 파일 μ‹œμŠ€ν…œ (File System)

도ꡬ μ„€λͺ… μ£Όμš” νŒŒλΌλ―Έν„°
glm_file_read 파일 λ‚΄μš© 읽기 path, encoding
glm_file_create μƒˆ 파일 생성 path, content, overwrite
glm_file_edit λ¬Έμžμ—΄ μΉ˜ν™˜ μˆ˜μ • path, old_string, new_string
glm_file_delete 파일 μ‚­μ œ (λ°±μ—… 보관) path
glm_file_rollback 이전 버전 볡원 path, version
glm_dir_list 디렉토리 λͺ©λ‘ path, recursive
glm_grep μ •κ·œμ‹ 파일 검색 pattern, path, case_sensitive

πŸ’» μ½”λ“œ μ‹€ν–‰ (Code Execution)

도ꡬ μ„€λͺ… μ£Όμš” νŒŒλΌλ―Έν„°
glm_code_run Python μƒŒλ“œλ°•μŠ€ μ‹€ν–‰ code, timeout
glm_shell_exec ν™”μ΄νŠΈλ¦¬μŠ€νŠΈ μ‰˜ μ‹€ν–‰ command, cwd

🌿 Git ν˜‘μ—… (Git Collaboration)

도ꡬ μ„€λͺ… μ£Όμš” νŒŒλΌλ―Έν„°
glm_git_status μ €μž₯μ†Œ μƒνƒœ 쑰회 repo_path
glm_git_commit 변경사항 컀밋 message, add_all
glm_git_log 컀밋 이λ ₯ 쑰회 n, oneline
glm_git_diff 변경사항 비ꡐ stat_only, commit

🌐 λ„€νŠΈμ›Œν¬ (Network)

도ꡬ μ„€λͺ… μ£Όμš” νŒŒλΌλ―Έν„°
glm_http_request HTTP μš”μ²­ (SSRF λ°©μ§€) url, method, body
glm_web_search DuckDuckGo μ›Ή 검색 query, max_results

🧠 λ©”λͺ¨λ¦¬ & 데이터 (Memory & Data)

도ꡬ μ„€λͺ… μ£Όμš” νŒŒλΌλ―Έν„°
glm_memory_save μž₯κΈ° λ©”λͺ¨λ¦¬ μ €μž₯ key, value, category
glm_memory_get λ©”λͺ¨λ¦¬ 쑰회 key
glm_memory_list 전체 λ©”λͺ¨λ¦¬ λͺ©λ‘ category, limit
glm_memory_delete λ©”λͺ¨λ¦¬ μ‚­μ œ key
glm_db_query SQLite 쿼리 μ‹€ν–‰ query, db_path, read_only

πŸ“Š 관리 & λ‘œκΉ… (Management)

도ꡬ μ„€λͺ… μ£Όμš” νŒŒλΌλ―Έν„°
glm_schedule_task μž‘μ—… μ˜ˆμ•½ (Cron) action, cron, command
glm_action_log μ—μ΄μ „νŠΈ ν™œλ™ 둜그 limit, action_filter

πŸ›‘οΈ λ³΄μ•ˆ μ•„ν‚€ν…μ²˜

이 ν”„λ‘œμ νŠΈλŠ” 4단계 λ³΄μ•ˆ 계측을 κ΅¬ν˜„ν•©λ‹ˆλ‹€:

flowchart TB
    subgraph "Layer 1: Sandbox"
        A[PROJECT_ROOT 경둜 검증]
    end
    
    subgraph "Layer 2: Network"
        B[SSRF λ°©μ§€ - 내뢀망 IP 차단]
        C[DNS Rebinding λ°©μ–΄]
    end
    
    subgraph "Layer 3: Execution"
        D[RCE λ°©μ§€ - ν™˜κ²½λ³€μˆ˜ 필터링]
        E[μ‰˜ ν™”μ΄νŠΈλ¦¬μŠ€νŠΈ]
    end
    
    subgraph "Layer 4: Data"
        F[μžλ™ λ°±μ—…]
        G[API ν‚€ 격리]
    end
    
    A --> B --> D --> F
보호 λŒ€μƒ μœ„ν˜‘ λ°©μ–΄ 쑰치
파일 μ‹œμŠ€ν…œ Path Traversal PROJECT_ROOT μ™ΈλΆ€ μ ‘κ·Ό 차단
λ„€νŠΈμ›Œν¬ SSRF 곡격 내뢀망 IP(10.x, 172.x, 192.168.x) 필터링
μ½”λ“œ μ‹€ν–‰ RCE / ν‚€ 유좜 ZHIPU_API_KEY λ“± 민감 ν™˜κ²½λ³€μˆ˜ 차단
μ‰˜ λͺ…λ Ή μ‹œμŠ€ν…œ 파괴 rm -rf, sudo λ“± μœ„ν—˜ λͺ…λ Ή 차단

[!WARNING] glm_shell_execλŠ” ν™”μ΄νŠΈλ¦¬μŠ€νŠΈμ— λ“±λ‘λœ μ•ˆμ „ν•œ λͺ…λ Ήλ§Œ μ‹€ν–‰ν•©λ‹ˆλ‹€. rm, sudo, chmod 777 등은 μ›μ²œ μ°¨λ‹¨λ©λ‹ˆλ‹€.


πŸ”§ ν™˜κ²½λ³€μˆ˜ μ„€μ •

λ³€μˆ˜λͺ… ν•„μˆ˜ μ„€λͺ… κΈ°λ³Έκ°’
ZHIPU_API_KEY βœ… GLM API 인증 ν‚€ -
PROJECT_ROOT βœ… μž‘μ—… λŒ€μƒ 디렉토리 ν˜„μž¬ 디렉토리
GLM_MODEL ❌ μ‚¬μš©ν•  λͺ¨λΈ glm-4-plus
GLM_BASE_URL ❌ API μ—”λ“œν¬μΈνŠΈ https://open.bigmodel.cn/api/paas/v4
GLM_TIMEOUT ❌ μš”μ²­ νƒ€μž„μ•„μ›ƒ (초) 120

πŸ“‚ ν”„λ‘œμ νŠΈ ꡬ쑰

antigravity_glm_mcp/
β”œβ”€β”€ πŸ“„ README.md                 # 이 λ¬Έμ„œ
β”œβ”€β”€ πŸ“„ requirements.txt          # Python μ˜μ‘΄μ„±
β”œβ”€β”€ πŸ“„ pyproject.toml            # νŒ¨ν‚€μ§€ 메타데이터
β”‚
β”œβ”€β”€ πŸ“ src/                      # μ†ŒμŠ€ μ½”λ“œ
β”‚   β”œβ”€β”€ server.py                # MCP μ„œλ²„ μ§„μž…μ 
β”‚   β”œβ”€β”€ models.py                # 곡톡 λͺ¨λΈ (ToolResponse λ“±)
β”‚   β”‚
β”‚   β”œβ”€β”€ πŸ“ core/                 # 핡심 인프라
β”‚   β”‚   β”œβ”€β”€ config.py            # μ„€μ • κ΄€λ¦¬μž
β”‚   β”‚   β”œβ”€β”€ glm_client.py        # GLM API ν΄λΌμ΄μ–ΈνŠΈ
β”‚   β”‚   β”œβ”€β”€ sandbox.py           # 경둜 λ³΄μ•ˆ 검증
β”‚   β”‚   └── backup.py            # μžλ™ λ°±μ—… μ‹œμŠ€ν…œ
β”‚   β”‚
β”‚   └── πŸ“ tools/                # 25개 도ꡬ κ΅¬ν˜„
β”‚       β”œβ”€β”€ glm_cmd.py           # μ§€λŠ₯ μœ„μž„ (cmd, bypass)
β”‚       β”œβ”€β”€ file_ops.py          # 파일 CRUD
β”‚       β”œβ”€β”€ dir_ops.py           # 디렉토리 μž‘μ—…
β”‚       β”œβ”€β”€ grep_ops.py          # 파일 검색
β”‚       β”œβ”€β”€ code_ops.py          # μ½”λ“œ μ‹€ν–‰
β”‚       β”œβ”€β”€ shell_ops.py         # μ‰˜ μ‹€ν–‰
β”‚       β”œβ”€β”€ git_ops.py           # Git ν˜‘μ—…
β”‚       β”œβ”€β”€ http_ops.py          # HTTP μš”μ²­
β”‚       β”œβ”€β”€ web_ops.py           # μ›Ή 검색
β”‚       β”œβ”€β”€ db_ops.py            # DB 쿼리
β”‚       β”œβ”€β”€ memory_ops.py        # λ©”λͺ¨λ¦¬ 관리
β”‚       β”œβ”€β”€ image_ops.py         # 이미지 뢄석
β”‚       β”œβ”€β”€ schedule_ops.py      # μž‘μ—… μ˜ˆμ•½
β”‚       └── reporting.py         # 둜그 관리
β”‚
β”œβ”€β”€ πŸ“ scripts/                  # μœ ν‹Έλ¦¬ν‹° 슀크립트
β”‚   └── install.py               # μžλ™ μ„€μΉ˜ 슀크립트
β”‚
β”œβ”€β”€ πŸ“ docs/                     # λ¬Έμ„œ
β”‚   β”œβ”€β”€ ARCHITECTURE.md          # μ•„ν‚€ν…μ²˜ 상세
β”‚   β”œβ”€β”€ TOOLS.md                 # 도ꡬ 레퍼런슀
β”‚   └── QUICKSTART.md            # λΉ λ₯Έ μ‹œμž‘ κ°€μ΄λ“œ
β”‚
β”œβ”€β”€ πŸ“ tests/                    # ν…ŒμŠ€νŠΈ μ½”λ“œ
β”‚   β”œβ”€β”€ local_tools_test.py      # 둜컬 도ꡬ 톡합 ν…ŒμŠ€νŠΈ
β”‚   └── simple_test.py           # μ§€λŠ₯ 도ꡬ ν…ŒμŠ€νŠΈ
β”‚
└── πŸ“ data/                     # λŸ°νƒ€μž„ 데이터 (Git μ œμ™Έ)
    β”œβ”€β”€ memory/                  # 영ꡬ λ©”λͺ¨λ¦¬ μ €μž₯μ†Œ
    └── action_logs.jsonl        # μ—μ΄μ „νŠΈ ν™œλ™ 둜그

πŸ§ͺ ν…ŒμŠ€νŠΈ

# κ°€μƒν™˜κ²½ ν™œμ„±ν™” ν›„

# 1. .env 파일 생성 (ꢌμž₯)
cp .env.sample .env
# .env νŒŒμΌμ„ μ—΄μ–΄ ZHIPU_API_KEYλ₯Ό μž…λ ₯ν•˜μ„Έμš”.

# 2. 둜컬 도ꡬ ν…ŒμŠ€νŠΈ (API ν‚€ λΆˆν•„μš”)
./.venv/bin/python tests/local_tools_test.py

# 3. μ§€λŠ₯ 도ꡬ ν…ŒμŠ€νŠΈ (API ν‚€ ν•„μš”)
# .env 파일이 μ—†λ‹€λ©΄ 직접 export ν•˜μ„Έμš”.
export ZHIPU_API_KEY="your-key" 
./.venv/bin/python tests/simple_test.py

πŸ“œ λΌμ΄μ„ μŠ€

이 ν”„λ‘œμ νŠΈλŠ” MIT License ν•˜μ— λ°°ν¬λ©λ‹ˆλ‹€. 자유둭게 μ‚¬μš©, μˆ˜μ •, λ°°ν¬ν•˜μ‹€ 수 μžˆμŠ΅λ‹ˆλ‹€.


<div align="center">

Made with ❀️ for Gemini Γ— GLM Collaboration

GitHub stars

</div>

Recommended Servers

playwright-mcp

playwright-mcp

A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.

Official
Featured
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.

Official
Featured
Local
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.

Official
Featured
Local
TypeScript
VeyraX MCP

VeyraX MCP

Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.

Official
Featured
Local
Kagi MCP Server

Kagi MCP Server

An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.

Official
Featured
Python
graphlit-mcp-server

graphlit-mcp-server

The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.

Official
Featured
TypeScript
Qdrant Server

Qdrant Server

This repository is an example of how to create a MCP server for Qdrant, a vector search engine.

Official
Featured
Neon Database

Neon Database

MCP server for interacting with Neon Management API and databases

Official
Featured
Exa Search

Exa Search

A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.

Official
Featured
E2B

E2B

Using MCP to run code via e2b.

Official
Featured