Prompt file imported from dav1dc-github-cs-org/autocomplete-to-agents (
.github/prompts/06-ide-instructions.prompt.md). Copyright stays with the author.
Operations want to catch shipments that are about to breach their service commitment.
Add a way to find shipments that are still in transit and have been in their
current status for longer than a threshold, and expose it on the CLI as
freightline at-risk --hours 48.
That is the whole brief. Everything else — where the code goes, what the query looks like, how errors are raised, how the tests are named, what gets a comment — you already know, because it is written down in this repository.
So before you write a line of code:
List every instruction file that applies to the files you are about to touch, and the specific rule from each that will change what you write. Quote the rule. Then implement it, and afterwards point at each place in your diff where one of those rules is visible in the output.
Finish with make check.
Then tell me the honest version: which of those rules would you have followed anyway, and which ones only happened because this repository wrote them down?