Use an AI spec to brief a coding agent

Work out a feature with an AI assistant, then pass the spec to a coding agent and verify the result by testing.

Done with ChatGPT· Intermediate· Feb 2026

What they did
They started by talking through the feature with ChatGPT until it produced a detailed spec. That document then went to a coding agent, which wrote the implementation and tests. Instead of inspecting the code line by line, they checked the behavior themselves by exercising the feature end to end.
How it went
They said this let them ship stronger applications more quickly. The assistant caught edge cases and standard practices they might have overlooked, while any bugs left behind were the sort they would likely have written themselves. The main risk was missing problems that only careful testing would reveal, especially with odd or unexpected input.
Tips
Treat your test suite as the real review step, especially for strange data and edge cases. If you rely on behavior checks instead of code reading, your tests need to be specific and thorough.

Try it yourself with Claude Code

Help me add [feature] to [project]. First ask me the questions you need until you can write a clear specification, then save that spec as SPEC.md. After that, implement the feature and add tests that verify the expected behavior. Before you delete files, change the database, or make any irreversible action, pause and ask for my approval. When you’re done, give me a short checklist I can use to try the feature myself.

Fill in the [brackets]. Agents can make mistakes, so review anything before it is sent, paid or deleted.

Read the original on news.ycombinator.com ↗

Get 3 agent uses like this every morning

With the day's top AI news. Free, and one click to unsubscribe.