Main content blocks
Section outline
In this discussion, you will begin building the AI agent you designed in Week 1. You will work in your individual server sandbox using your Moodle installation, terminal, and AgentLab.
Your goal this week is to move from an idea to a working first version of your agent. Your agent does not need to be complete. It should demonstrate that the basic components are beginning to work and that you understand how your agent will connect with Moodle.
Because participants may be developing different kinds of agents, your technical implementation may be different from that of other participants. AgentLab will help you determine an appropriate development approach based on what you want your agent to accomplish and the Moodle version installed in your sandbox.
What You Will Build
You will create the first working version of your Moodle AI agent. Your first version should have a clearly defined purpose, operate within or connect with your Moodle environment, have a basic way of receiving a request, connect with an AI model when required by your design, process a test request, return a meaningful response, and provide a foundation that you can make more agentic in Week 3.
Before You Begin
Confirm that you can access your individual server sandbox, your Moodle installation, Moodle Site administration, the terminal, and AgentLab. Also identify the Moodle version installed in your sandbox.
Do not share passwords, API keys, server credentials, database credentials, or other private information.
Step 1: Review the Agent You Defined in Week 1
Return to the AI agent you defined during Week 1. Review the problem your agent will address, who will use it, what the agent should understand, what the agent should help the user accomplish, and how it could interact with Moodle.
Your project can change as you develop it. You do not need to follow your original idea if testing leads you in a better direction.
Step 2: Begin With AgentLab
Open the Moodle AgentLab and describe the agent you want to build. Provide relevant information about your development environment.
You can begin with:
"I want to build an AI agent for Moodle that helps learners identify what they should work on next. I have my own server sandbox, terminal access, Moodle administrator access, and Moodle [version]. Help me determine the best architecture and guide me through building a first working version."
Do not ask AgentLab to build everything at once. Work through the project in stages so you understand what is being created and why.
Step 3: Plan the Architecture
With AgentLab, determine how your agent should be structured. Depending on the purpose of your agent, this might involve a Moodle plugin, an AI service, Moodle APIs, an external service, or a combination of components.
Your architecture does not have to be identical to another participant's architecture.
Before creating files, make sure you can explain where the agent will operate, how a user will interact with it, what information the agent needs, how it will communicate with an AI model if one is required, and how it will connect with Moodle.
Step 4: Create the Core Components
Use your terminal and Moodle environment to begin creating the components identified in your plan.
Moodle AgentLab can guide you through creating files, directories, code, Moodle components, and configuration appropriate to your project.
Before running commands or adding code, read AgentLab's explanation so you understand what the step is intended to accomplish. Build incrementally rather than attempting to create the complete agent at once.
Step 5: Connect an AI Model
If your agent requires a generative AI model, connect it to the AI provider you have selected for your project. For example, your agent may use Gemini, ChatGPT, or another AI model available to you.
Keep API credentials secure. Never place API keys in public code, Moodle Additional HTML, screenshots, forum posts, or other locations accessible to participants.
Ask AgentLab to help you determine an appropriate way to store credentials securely in your particular environment.
Step 6: Connect the Agent With Moodle
Begin connecting your agent with the Moodle environment. The type of connection will depend on the purpose of your agent.
For example, an agent might eventually work with course information, activities, resources, completion information, navigation, teacher support, or other Moodle functionality.
At this stage, the connection can be simple. More advanced Moodle context and agentic reasoning will be developed in Week 3.
Step 7: Test the First Version
Test your agent with a simple request related to its intended purpose. Confirm that the request reaches the appropriate component and that the agent returns a response.
Do not worry if the response is not yet sophisticated. The goal is to establish a working foundation.
Test several requests and note what works and what still needs improvement.
Step 8: Troubleshoot With Moodle AgentLab
If something does not work, use Moodle AgentLab to help investigate the problem. Provide the relevant error message, describe what you expected to happen, and explain what actually happened.
Do not share passwords, API keys, or other credentials with AgentLab when troubleshooting.
Rather than immediately replacing large amounts of code, try to identify the source of the problem and make one change at a time.
Step 9: Create Your Video Demonstration
Create a short screen recording using the recording tool of your choice.
Your video should include your chat with Moodle AgentLab and a brief introduction to your AI agent, the teaching or learning problem your agent addresses, your Moodle site, the first working version of your agent, a demonstration of a request and response, a brief explanation of how the agent currently works, how Moodle AgentLab supported your development, one challenge you encountered and how you addressed it or a challenge for which you still need help, and what you plan to develop next.
Do not display passwords, API keys, server credentials, database information, or other sensitive information in your recording.
Step 10: Share Your Development Progress
- Upload your video to Google Drive.
- Set the sharing permission to Anyone with the link can view.
- Copy the Google Drive link and add it to your discussion post.
- Create a post titled My Moodle AI Agent: Your Name.
- Briefly describe what your agent currently does and include the link to your video.
Your agent does not need to be finished. The purpose of this discussion is to demonstrate your development process and first working version.
Completion Statement
End your post with:
My Moodle AI agent has a working foundation and is ready to become more agentic.
Peer Reply
Reply to at least one participant. Watch their video, comment on one aspect of their agent, ask one relevant question, and offer one helpful suggestion.
Your classmates may be developing very different agents. Use the discussion to learn from the different approaches, architectures, problems, and solutions being explored.
- Receive a grade
- Receive a passing grade
| Status | Discussion | Started by | Last post | Replies | Actions |
|---|---|---|---|---|---|
|
Locked
|
|
|
4 |
|