What is the primary goal of the custom prompt being built?	To help users write code specific for AWS use cases.
What are the three specific output types the prompt must generate?	Python, JSON configuration, or a raw regular expression.
What is the critical constraint on the prompt's output format?	It must provide only the code/output without any explanation, header, footer, or other text.
What are the two main steps in building the prompt evaluation workflow?	Writing the prompt and assembling a dataset.
How is the input dataset structured for this evaluation?	An array of JSON objects, where each object contains a `task` property.
What model is suggested for generating the evaluation dataset to improve speed?	Haiku.
What technique is used to extract JSON from Claude's text response?	Using a combination of a pre-filled assistant message and a stop sequence (e.g., three backticks).
What Python module function is used to convert the raw text response into a usable data structure?	`json.loads()`.
What is the purpose of the `GenerateDataSet` helper function?	To prompt Claude to automatically generate the evaluation test set (the tasks).
What is the final step after generating the dataset?	Saving the dataset into a JSON file.