Codeaid

Inspiration

As a CLI-based developer, I wanted a utility that would support development-level activities, such as code summarization, refactoring, defect or security issue identification, etc.

What it does

A command-line utility that takes a source file and performs a review of the code based upon a given goal (summarize, defects, complexity, refactor, unit-test, etc.).

How we built it

Developed iteratively using the script itself. At each iteration, the codeaid utility was applied to ensure no defects, security-issues, and best structure with minimal complexity. Developed with Python and OpenAI.

Challenges we ran into

Developed a PromptManager class that encapsulates the most useful prompts for LLMs and identifies them by a goal. This was a key in the development, and ended up being a key part of the delivery.

Accomplishments that we're proud of

Building a tool that guided that development of the tool to a successful end.

What we learned

Python is amazing as a prototyping tool. OpenAI easily integrates into Python, making it simple to use. Prompts matter, how they're structured, and persona-based prompts are key in guiding LLMs to the best responses.

What's next for Codeaid

What's next are new prompts as my journey as a developer with LLMs continues.

Built With

• openai
• python