AI-Powered Development
Overview
ShadcnSpace Dashboard is designed to be AI-First. It includes a comprehensive system of instructions and prompts that allow AI agents (like Claude 3.5 Sonnet, GPT-4, or Gemini) to understand the codebase and perform complex tasks autonomously.
1. AGENTS.md (System Instructions)
The AGENTS.md file located at the root of the
project is the master instruction set for any AI agent. It
contains:
- Role & Context: Defines the AI as a Frontend Architect.
- Project Overview: Explains the architecture and core features.
- Tech Stack: Lists all libraries and frameworks used.
- Key Locations: Maps out the file structure for quick navigation.
- Naming Conventions: Ensuring the AI writes code that matches the project style.
AGENTS.md to
your AI agent at the start of a session to give it full
context of the project.
For more details about AGENTS.md: https://agents.md/
2. AI Prompt Library
The /prompts directory is organized into
functional categories. Below is the complete list of available
prompts across both Starterkit and
Main
versions of the project.
Explore Prompts by Category
Apps
Move specific modules or pages into standalone applications for micro-frontend setups.
Auth
Generate standardized authentication flows including login, registration, and recovery pages.
Branding
Globally update brand identity, including logos, company names, and project-wide metadata.
Components
Generate advanced TanStack tables with filtering, sorting, and pagination using shadcn/ui.
Integration
Exclusive prompt for Starterkit to migrate any feature, component, or logic from the Main project.
Dashboard
Build complex layout blueprints and modify existing data sources for dashboard widgets and charts.
Forms
Generate React Hook Forms with integrated Zod validation and consistent field styling.
Language
Implement full i18n support, manage switches, and decommission existing language locales.
Navigation
Manage site structure: add pages, configuration menus, nesting, and dynamic badges.
Theming
Control dark/light modes, modify color variables, and configure RTL/LTR direction support.
3. How to Create a New Prompt
To create a new automated task, follow the structure of existing prompts:
# Template for a new Prompt
You are a [Role]. Your task is to [Primary Goal].
## Step 1: [Action]
**Target Files:** [Path to files]
**Instructions:** [Detailed steps for the AI]
## Verification Checklist
- [ ] [Verification step 1]
- [ ] [Verification step 2]
4. Recommended Workflow
To get the best results from AI when working with shadcnspace-dashboard Pro, we recommend following this standardized four-step cycle:
Initialize Context
Start your session by uploading or pasting the contents
of AGENTS.md. This gives the AI full
context of the project architecture, tech stack, and
naming conventions.
Select & Deploy Prompt
Choose the relevant .prompt.md file from
the /prompts directory. Copy its contents
and paste it into the AI chat to define the specific
task.
Collaborative Execution
Provide the AI with your specific requirements (e.g., brand name, data fields). The AI will process the instructions and generate the necessary file changes or new components.
Verify & Finalize
Use the Verification Checklist provided
at the end of each prompt. Run
npm run dev and
npm run lint to ensure the code is
error-free and matches the design.