System Prompts MCP Server

1# 📝 System Prompts MCP Server
2 
3> **Access system prompts from AI tools in your workflow.** Browse and fetch prompts from Devin, Cursor, Claude, GPT, and more. Model-aware suggestions help you find the perfect prompt for your LLM.
4 
5An [MCP (Model Context Protocol)](https://modelcontextprotocol.io) server that exposes a collection of system prompts, summaries, and tool definitions from popular AI tools as MCP tools for AI coding environments like Cursor and Claude Desktop.
6 
7## Why Use System Prompts MCP?
8 
9- 🔍 **Automatic Discovery** – Every prompt in `prompts/` is automatically exposed as an MCP tool
10- 🎯 **Model-Aware Suggestions** – Get prompt recommendations based on your LLM (Claude, GPT, Gemini, etc.)
11- 📚 **Comprehensive Collection** – Access prompts from Devin, Cursor, Claude, GPT, and more
12- 🚀 **Easy Setup** – One-click install in Cursor or simple manual setup
13- 🔧 **Extensible** – Add your own prompts and they're automatically available
14 
15## Quick Start
16 
17Ready to explore system prompts? Install in seconds:
18 
19**Install in Cursor (Recommended):**
20 
21[🔗 Install in Cursor](cursor://anysphere.cursor-deeplink/mcp/install?name=system-prompts-mcp&config=eyJzeXN0ZW0tcHJvbXB0cy1tY3AiOnsiY29tbWFuZCI6Im5weCIsImFyZ3MiOlsiLXkiLCJzeXN0ZW0tcHJvbXB0cy1tY3Atc2VydmVyIl19fQ==)
22 
23**Or install manually:**
24 
25```bash
26npm install -g system-prompts-mcp-server
27# Or from source:
28git clone https://github.com/JamesANZ/system-prompts-and-models-of-ai-tools.git
29cd system-prompts-and-models-of-ai-tools && npm install && npm run build
30```
31 
32## Features
33 
34### Core Tools
35- **`list_prompts`** – Browse available prompts with filters (service, flavor, provider)
36- **`get_prompt_suggestion`** – Get ranked prompt suggestions for your LLM and keywords
37- **`<service>-<variant>-<flavor>`** – Direct access to any prompt (e.g., `cursor-agent-system`, `devin-summary`)
38 
39### Automatic Discovery
40- Scans `prompts/` directory for `.txt`, `.md`, `.yaml`, `.yml`, `.json` files
41- Each file becomes a dedicated MCP tool
42- Infers metadata (service, variant, LLM family, persona hints)
43 
44### Persona Activation
45- Each tool call includes a reminder for the model to embody the loaded prompt
46- Helps models behave like the original service (Devin, Cursor, etc.)
47 
48## Installation
49 
50### Cursor (One-Click)
51 
52Click the install link above or use:
53 
54```
55cursor://anysphere.cursor-deeplink/mcp/install?name=system-prompts-mcp&config=eyJzeXN0ZW0tcHJvbXB0cy1tY3AiOnsiY29tbWFuZCI6Im5weCIsImFyZ3MiOlsiLXkiLCJzeXN0ZW0tcHJvbXB0cy1tY3Atc2VydmVyIl19fQ==
56```
57 
58### Manual Installation
59 
60**Requirements:** Node.js 18+ and npm
61 
62```bash
63# Clone and build
64git clone https://github.com/JamesANZ/system-prompts-and-models-of-ai-tools.git
65cd system-prompts-and-models-of-ai-tools
66npm install
67npm run build
68 
69# Run server
70npm start
71```
72 
73### Claude Desktop
74 
75Add to `claude_desktop_config.json`:
76 
77**macOS**: `~/Library/Application Support/Claude/claude_desktop_config.json`  
78**Windows**: `%APPDATA%\Claude\claude_desktop_config.json`
79 
80```json
81{
82  "mcpServers": {
83    "system-prompts-mcp": {
84      "command": "node",
85      "args": ["/absolute/path/to/system-prompts-and-models-of-ai-tools/dist/index.js"],
86      "env": {
87        "PROMPT_LIBRARY_ROOT": "/absolute/path/to/system-prompts-and-models-of-ai-tools/prompts"
88      }
89    }
90  }
91}
92```
93 
94Restart Claude Desktop after configuration.
95 
96## Usage Examples
97 
98### List Available Prompts
99Browse prompts with optional filters:
100 
101```json
102{
103  "tool": "list_prompts",
104  "arguments": {
105    "service": "cursor",
106    "flavor": "system",
107    "limit": 10
108  }
109}
110```
111 
112### Get Prompt Suggestions
113Find the best prompt for your LLM and use case:
114 
115```json
116{
117  "tool": "get_prompt_suggestion",
118  "arguments": {
119    "userLlm": "claude-3.5-sonnet",
120    "keywords": ["code", "pair programming"]
121  }
122}
123```
124 
125### Access a Specific Prompt
126Call a prompt directly by its tool name:
127 
128```json
129{
130  "tool": "cursor-agent-system",
131  "arguments": {}
132}
133```
134 
135Get structured metadata only:
136 
137```json
138{
139  "tool": "cursor-agent-system",
140  "arguments": {
141    "format": "json"
142  }
143}
144```
145 
146## Adding Your Own Prompts
147 
148Add prompts by placing files in the `prompts/` directory:
149 
150**Supported formats:** `.txt`, `.md`, `.yaml`, `.yml`, `.json`
151 
152**Directory structure:**
153```
154prompts/My Service/
155  ├── System Prompt.txt     → Tool: "my-service-system-prompt-system"
156  └── tools.json            → Tool: "my-service-tools-tools"
157```
158 
159- Directory names become the service name
160- File names create tool variants
161- Files are automatically classified as system prompts, tools, or summaries
162 
163After adding prompts, restart the MCP server. Use `list_prompts` to find your custom prompts.
164 
165**Custom directory:** Set `PROMPT_LIBRARY_ROOT` environment variable to use a different location.
166 
167## Use Cases
168 
169- **AI Tool Developers** – Reference and adapt prompts from successful AI tools
170- **Researchers** – Study how different tools structure their system prompts
171- **Developers** – Find the perfect prompt for your LLM and use case
172- **Prompt Engineers** – Compare and learn from proven prompt patterns
173 
174## Technical Details
175 
176**Built with:** Node.js, TypeScript, MCP SDK  
177**Dependencies:** `@modelcontextprotocol/sdk`, `zod`  
178**Platforms:** macOS, Windows, Linux
179 
180**Environment Variables:**
181- `PROMPT_LIBRARY_ROOT` (optional): Override prompt root directory (defaults to `prompts/`)
182 
183**Project Structure:**
184- `src/` – TypeScript MCP server implementation
185- `dist/` – Compiled JavaScript
186- `prompts/` – Prompt library and original documentation
187 
188## Contributing
189 
190⭐ **If this project helps you, please star it on GitHub!** ⭐
191 
192Contributions welcome! Feel free to adapt the discovery logic, add tests, or extend metadata inference for new prompt formats.
193 
194## License
195 
196See the original repository for license information.
197 
198## Support
199 
200If you find this project useful, consider supporting it:
201 
202**⚡ Lightning Network**
203```
204lnbc1pjhhsqepp5mjgwnvg0z53shm22hfe9us289lnaqkwv8rn2s0rtekg5vvj56xnqdqqcqzzsxqyz5vqsp5gu6vh9hyp94c7t3tkpqrp2r059t4vrw7ps78a4n0a2u52678c7yq9qyyssq7zcferywka50wcy75skjfrdrk930cuyx24rg55cwfuzxs49rc9c53mpz6zug5y2544pt8y9jflnq0ltlha26ed846jh0y7n4gm8jd3qqaautqa
205```
206 
207**₿ Bitcoin**: [bc1ptzvr93pn959xq4et6sqzpfnkk2args22ewv5u2th4ps7hshfaqrshe0xtp](https://mempool.space/address/bc1ptzvr93pn959xq4et6sqzpfnkk2args22ewv5u2th4ps7hshfaqrshe0xtp)
208 
209**Ξ Ethereum/EVM**: [0x42ea529282DDE0AA87B42d9E83316eb23FE62c3f](https://etherscan.io/address/0x42ea529282DDE0AA87B42d9E83316eb23FE62c3f)
210