HomeProjectsArticlesPK BoardAbout
Explore Projects

System Prompts and Models of AI Tools

Active
GitHub Unknown GPL-3.0

Description

Full system prompts, internal tools and AI models from 40+ popular AI tools including Cursor, Devin, Windsurf, Manus, Lovable, and more.

Key Features

  • System prompt collection — Complete system prompts from 40+ AI tools (Cursor, Devin, Windsurf, Manus, etc.)
  • AI model intelligence — Records internal AI model names and versions used by each tool for tech stack insights
  • Internal tool definitions — Exposes registered function/tool schemas revealing agent capabilities
  • Continuous updates — Prompt content updated as tools release new versions, keeping information current
  • Security research reference — Empirical data for studying prompt injection and system prompt extraction risks

Use Cases

💡 Study competitor AI products' prompt engineering strategies and learn industry-leading prompt design patterns
💡 Analyze internal tool definitions of mainstream AI coding tools for reference in your own agent development
💡 Conduct AI security audits to evaluate system prompt leakage risks and prompt injection defenses
💡 Understand underlying models used by AI tools to inform technology selection and model comparison decisions
💡 Use as a prompt engineering learning resource to improve prompt quality in your own AI applications

Strengths & Limitations

✅ Strengths

  • • Actively maintained, recent updates
  • • High community interest (143.3k stars)
  • • Responsive to issues, low backlog

⚠️ Limitations

  • • Restrictive license (GPL-3.0)

Tags

prompt-engineering agent tools llm api

Categories

🔌 MCP Tools

Quick Start

# Clone the repo to browse system prompts
git clone https://github.com/x1xhlol/system-prompts-and-models-of-ai-tools.git
cd system-prompts-and-models-of-ai-tools

# Browse directory structure, one folder per tool
ls -la

# View a specific tool's prompt (e.g. Cursor)
# cat Cursor/system_prompt.md
Visit GitHub

Related Projects

Gibberlink

4.9k · TypeScript
Stale B

A protocol where two conversational AI agents switch from English to a sound-level protocol after confirming they are both AI agents, improving inter-agent communication efficiency.

typescriptagentapi +2
  • · AI-to-AI agents switch from English to ggwave sound-level protocol automatically
  • · Won 1st place at 11labs x a16z international hackathon, covered by Forbes and TechCrunch
  • · Built on ggwave data-over-sound library with ElevenLabs conversational AI integration
View on GitHub

Apify MCP Server

5.8k · TypeScript
Active A

MCP server that enables AI agents to extract data from social media, search engines, maps, and e-commerce sites using thousands of Apify scrapers.

typescriptmcptools +3
  • · Access to thousands of ready-made scrapers, crawlers, and automation tools from Apify Store
  • · OAuth support for one-URL connection from Claude.ai, VS Code, Cursor, and other MCP clients
  • · Hosted server at mcp.apify.com with Streamable HTTP transport (SSE deprecated April 2026)
View full details → View on GitHub

Google Workspace CLI

30.7k · Rust
Active A

A unified CLI for Google Workspace covering Drive, Gmail, Calendar, Sheets, Docs and more, with built-in AI agent skills for automation.

cliagentrust +2
  • · Dynamic command discovery — Reads Google Discovery Service at runtime to auto-build all API commands; new Google endpoints auto-supported
  • · 40+ agent skills — Built-in AI agent skills with structured JSON output for LLMs to manage Workspace directly
  • · Multiple auth methods — Supports interactive OAuth, Service Account, pre-obtained tokens, CI/headless credential export
View full details → View on GitHub