Prompting Techniques
Why Use Advanced Prompting?
Effective prompting helps you get the most out of Monk in various situations:
- Troubleshooting: When you encounter errors or unexpected behavior in your deployments
- Decision-making: When you need guidance on infrastructure choices or best practices
- Efficiency: When you want to complete multiple tasks with minimal interaction
- Team collaboration: When working with both Monk and other assistants like Copilot
Asking Direct Questions
Direct questions help when you need specific information or guidance:
@monk explain the error to me@monk what can I do in this situation?@monk how would you size infrastructure for this project?
These prompts are ideal when you’re stuck, troubleshooting, or need expert infrastructure advice.
Chaining Commands
Chain several instructions together to accomplish complex tasks with a single prompt:
@monk deploy the project on GCP and open preview in the browser@monk check the logs and restart containers if there is an error
This technique shines when you have a sequence of related tasks to complete or when setting up automated workflows.
Monk-to-Copilot Handoff
When Monk suggests code changes:
- Delete the
@monk
mention - Ask Copilot:
@workspace do what monk said
This workflow is perfect when you need infrastructure expertise from Monk but want Copilot to implement the changes in your codebase.
Copilot-to-Monk Delegation
Have Copilot delegate tasks to Monk:
@workspace ask monk to deploy this
Use this approach when you’re primarily working with Copilot but need Monk’s deployment capabilities.
Context Sharing
Provide Monk with relevant context to get more accurate assistance:
@monk Use a t2.micro EC2 instance. Would this be sufficient for this app?
Iterative Refinement
Start with a general request and then refine based on Monk’s response:
@monk how should I deploy this Astro app?// After Monk responds@monk let's go with Netlify. Can you use my existing Netlify account?