🤖 AI Prompts & Custom Skills Catalog

Welcome to the AI Prompts & Custom Skills directory. Konture is built from the ground up to be AI-agent friendly, offering structured, high-context, and API-accurate prompt assets.

By providing these prompts to modern AI coding assistants (such as Gemini Advanced, Claude Pro, ChatGPT) or loading them as system instructions in AI-integrated IDEs (such as Cursor, Windsurf, GitHub Copilot), you can automate architectural test setup, test generation, and review processes with zero syntax errors or hallucinations.


📐 Catalog of AI Prompts & Skills

Prompt / Skill File Description Target Use Case Executable Agent Skill
🤖 AI Onboarding & Setup Directs an AI agent to inspect the current codebase structure (modules, catalog config) and install/configure a dedicated test module. Automated installation, :konture-test module creation, and CI wiring. 🤖 setup-konture
✍️ Unified AI Test Writing & Extensible Guardrails Directs an AI assistant to systematically build, extend, or review architecture rules using extensible guardrails and compile-safe DSL API references. Designing and extending large-scale, syntax-accurate Konture architecture tests. 📐 konture-architecture-tests

📂 How to Open and Use These Prompts

1. Traditional Chat Assistants (Gemini, Claude, ChatGPT Web)

  1. Open any of the prompt markdown files above in your IDE or directly on GitHub.
  2. Copy the content from the ## 📋 Copy-Pasteable System Prompt section of the file.
  3. Paste the prompt as the initial context at the start of your chat session.
  4. Add your custom instructions (e.g., “Set up Konture in this multi-module Android project” or “Write a test to ensure our presentation layer doesn’t leak into the database layer”).

2. AI-Integrated IDEs (Cursor, Windsurf, Copilot, etc.)

  • System Instructions / AI Rules: Copy the content of the prompt and append it to your workspace’s .cursorrules, .windsurfrules, or Copilot instruction settings.
  • Direct File Referencing (@ mentions): Use the reference features (e.g., typing @docs/ai-prompts/writing-tests-prompt.md in Cursor/Windsurf) to instantly feed the complete API reference list directly into the model’s active context window.

🔌 What are “Custom Skills” for Autonomous Agents?

If you are using fully autonomous coding agents (such as Google Antigravity, Cursor Composer, or other MCP-compatible systems), this repository contains pre-configured Custom Skill Packages in the root skills/ directory:

  • Auto-Discovery: Agentic tools automatically scan and discover files in customization roots (e.g., skills/ or .agents/skills/).
  • Tool Equipping: When you ask an agent to “Set up architecture tests” or “Check our layer boundaries,” the agent discovers these files, registers them as “skills,” and loads the detailed markdown instructions into its background system prompts.
  • Frictionless Action: Equipped with these rules, autonomous agents can safely run inspections, edit Gradle builds, create modules, and compile tests without needing your manual oversight at each step.

[!TIP]

💡 Pro-Tip: Zero-Hallucination Testing

When asking AI tools to write tests, always mention: “Please check @writing-tests-prompt.md first to reconcile actual Konture DSL signatures rather than generating ArchUnit-style pseudo-code.” This saves time and ensures your generated code compiles perfectly on the first run.


This site uses Just the Docs, a documentation theme for Jekyll.