Claude Code Integration Guide

---

1. Introduction

informat is a system-level Skill built for the Informat AI Low-Code Development Platform, enabling natural language invocation of Informat capabilities within Claude Code or other AI agents, including:

• Data table creation and management
• Automation workflow configuration
• Dashboard building
• Workflow (BPMN) design
• Script execution
• Data read/write operations

Core capability: Let AI directly operate the low-code platform

---

2. Download Skill

Download the Skill files: informat-2.36.zip

informat/
├── skill.yaml
├── tools/
│   ├── informat.md
│   └── references/
├── scripts/
│   ├── call.js
│   └── .env

---

3. Connect to Claude Code

Place the Skill Directory

Place the extracted informat folder into the Claude Code configuration directory:

your-project-directory/.claude/skills/informat/

Reference: Extending Claude with Skills

Obtain Informat ApiKey

1. Log in to the Informat platform
2. Visit {host}/workbench/account/apikey
3. Click "New ApiKey" and set an expiration time
4. Copy the generated ApiKey

Configure Environment Variables

Create or edit a .env file in the scripts/ directory with the following content:

INFORMAT_HOST=<informat host>
INFORMAT_AGENT_TOKEN=<your-agent-token>

Parameter	Description
INFORMAT_HOST	Informat server address
INFORMAT_AGENT_TOKEN	Agent access token

Enable the Skill

Restart Claude Code or reload skills. The following prompt in the conversation indicates successful loading:

🛠️ Skill loaded: informat (Informat AI Low-Code Development Platform)

---

4. Capability Overview

The current Skill supports 119 system methods, covering the following main capability modules:

Module	Capabilities
Data Tables	Table, field, and record management
Automation	Automation workflows
Workflows	BPMN processes
Dashboards	Charts and analytics
APIs	Interface management
Scripts	JavaScript execution
Scheduled Tasks	Scheduled jobs
Websites	Page resource management
Internationalization	Multi-language support

Use Cases

• AI-powered CRM / ERP system building
• Automated business management system generation
• Data analytics dashboard automation
• AI-driven low-code development

---

5. Example: Auto-generate a System via Prompt

Below is a typical example: using a natural language prompt to drive Claude Code + Informat to automatically build a complete project management system.

Prompt Input

# Project Management System (PMS)

## 1. Product Overview

**Goal**: Provide an end-to-end solution for managing the full lifecycle of software projects, including project planning, execution, tracking, and delivery.

---

## 2. Feature Structure

- Project Management
- Member Management
- Milestone Management
- Task Management
- Defect Tracking
- Progress & Data Analytics

---

## 3. Feature Modules

### 3.1 Project Management

- Create / Edit / Delete projects
- Project dashboard (progress, risks, defects)

### 3.2 Member Management

- Add / Remove members
- Assign roles
- Workload tracking

### 3.3 Milestone Management

- Define milestones
- Link tasks and defects
- Status tracking

### 3.4 Task Management

- Task creation and assignment
- Subtasks and dependencies
- Kanban / Gantt chart / Calendar views
- Time tracking

### 3.5 Defect Tracking

- Report / Assign / Fix / Close defects
- Severity and priority levels
- Defect analytics