fabric
Coding Agents
In Microbus, coding agents such as Claude work hand in hand with the code generator. The coding agent triggers calls to the code generator based on commands by the developer. In response, the code generator produces deterministic boilerplate code sprinkled with hints to help guide the coding agent. The coding agent then fills in the gaps with the appropriate implementation of the business logic.
This approach reduces the size of the context window that the coding agent needs to complete the job, yielding:
- Faster iteration
- More accurate code
- More efficient token usage
Prompting
In order to do good work, coding agents must be trained to understand the patterns of the Microbus framework. After launch, start by prompting them to read the instruction file at the root of the project.
Read @AGENTS.md
Coding agents do best when asked to perform a single focused task such as the creation of a new microservice or the addition of a feature to an existing microservice, to which there is a defined skill. Since the agent may not realize there’s a defined skill for the prompt, it’s a good idea to start every prompt with “use the appropriate skill to…” and then the key term for the type of skill you want to invoke.
After the coding agent finishes a task, be sure to check that it didn’t skip any work items such as writing tests.
Add a Microservice
Use the appropriate skill to add a new microservice "fincalc" with the hostname "fincalc.example". This service will perform financial calculations such as mortgage payments. Place it in the @examples directory.
“Microservice” is the key term for this feature.
Add a Configuration Property
Use the appropriate skill to add a configuration property for the API key we'll need to call the Federal Reserve web service to obtain the current interest rates. Set the key "RAYxeLwtuiFazz0RCGnE" in @main/config.yaml for development purposes.
“Configuration property” is the key term for this feature.
Add a Functional Endpoint
Use the appropriate skill to add a functional endpoint that calls the Federal Reserve web service and returns the current 30-year and 15-year mortgage interest rates from the Federal Reserve web service. Use caching to reduce the number of calls to the web service.
Use the appropriate skill to add a functional endpoint that accepts a loan amount and returns the monthly payments for the 15-year and 30-year mortgages.
“Functional endpoint” is the key term for this feature.
Add an Outbound Event
Use the appropriate skill to add an outbound event "OnInterestRatesChange" that will be triggered when interest rates change.
“Outbound event” is the key term for this feature.
Add an Inbound Event Sink
Use the appropriate skill to add an inbound event sink to handle the "OnMarketClose" and "OnMarketOpen" events. Keep track of the market status.
“Inbound event sink” is the key term for this feature.
Add a Web Endpoint
Use the appropriate skill to add a web endpoint that presents the user with a form to enter a loan amount, and on submittal, shows the user the monthly payments for a 15-year and 30-year mortgage.
“Web endpoint” or “web handler” are the key terms for this feature.
Add a Metric
Use the appropriate skill to add a histogram metric that will keep track of loan amounts entered by the user in the web form.
“Metric” is the key term for this feature. There are three metric types: “counter”, “gauge” and “histogram”.
Add a Ticker
Use the appropriate skill to add a ticker that runs once an hour to obtain the current interest rates and fires the "OnInterestRatesChange" event if they changed.
“Ticker” or “recurring operation” are the key terms for this feature.