This plugin provides a set of agent tools for automatically generating Dataiku plugins based on user objectives. The tools work together to research documentation, generate code, and deploy plugins to GitHub and Dataiku instances.

1. Documentation Research Tool: Searches the web for relevant documentation about the plugin you want to create. It can retrieve information from various sources, including Dataiku documentation, GitHub repositories, and developer blogs.

2. Plugin Code Generation Tool: Generates code for your Dataiku plugin based on your objectives and the researched documentation. Creates a fully structured Dataiku plugin including proper configuration files, Python code, and requirements.

3. GitHub Integration Tool: Pushes the generated plugin code to a GitHub repository, either creating a new repository or updating an existing one.

4. Dataiku API Tool: Integrates with the Dataiku REST API to create plugins in a Dataiku instance and link them to Git repositories. This enables end-to-end deployment from code generation to installation.

5. Airtable Ticket Tool: Creates tickets in Airtable for tracking feature requests, bugs, or support inquiries. This can be used when a plugin generation request requires additional review or specialized development.

The typical workflow for using these tools is:

1. Define your plugin objective and requirements
2. Use the Documentation Research Tool to gather relevant information
3. Use the Plugin Code Generation Tool to create the plugin code
4. Use the GitHub Integration Tool to publish your plugin to GitHub
5. Use the Dataiku API Tool to create the plugin in your Dataiku instance and link it to the GitHub repository
6. Use the Airtable Ticket Tool if the plugin requires additional review or customization
7. Test and refine the generated plugin

• Dataiku Data Science Studio (DSS) 13.4 or later
• Google Custom Search API key for documentation research
• OpenAI API key for code generation
• GitHub account and personal access token for repository integration
• Dataiku API access for plugin deployment
• Airtable API key for ticket creation

1. Download the plugin
2. Install it in your Dataiku instance
3. Configure the required API connections

An AI agent would typically use these tools in sequence:

1. First, it would use the Documentation Research Tool to search for information about creating specific types of Dataiku plugins
2. Next, it would use the Plugin Code Generation Tool to generate code based on the research and user objectives
3. Then, it would use the GitHub Integration Tool to publish the generated code to a GitHub repository
4. Finally, it would use the Dataiku API Tool to create a plugin in the Dataiku instance and link it to the GitHub repository

If the request is beyond the agent's capability, it can create a ticket in Airtable using the Airtable Ticket Tool for human review.

Copyright (c) 2025 Dataiku SAS