HackStack

Project Info

GitScribe

Devpost

Inspiration

Maintaining quality documentation is a time consuming and difficult task and is also prone to falsification and misrepresentation which is why we wanted to automate these processes by leveraging Anthropic's AI capabilities.

What it does

Our tool uses Claude to analyze changes in a repo and create commit messages and documentation that accurately reflect the content and impact of the changes. By inputting the repo url our tool also analyzes all contributions to highlight top contributors based on the number of quality commits.

How we built it

We used the MERN stack to develop the website, and utilized the Agentic SDK to handle our LLM API calls. The commit cli was developed in Python which also interfaced with MongoDB, our database, and anthropic's claude to produce commit messages, code evals, and comprehensive documentation.

Challenges we ran into

We were having trouble managing the database in a way that didn't interfere with the various API calls that we had through the project. We also had trouble getting structured output from Claude API calls. However, we were eventually able to achieve consistent structured output through a workaround. Same issues plagued the cli until a way to manage api_keys properly was found.

Accomplishments we're proud of

This is a tool we would actually use in our workflow and we learned new skills that will contribute to our professional development

What we learned

For some of our team we learned how to create command line interfaces and manage APIs, for others we learned how to use React and front-end development principles

What's next

for GitGenie Making it scalable for larger repos, and optimization of our algorithms and data structures . Additionally, it would be cool to create custom classification network or a custom AI agent to better classify commits.

Analysis

Compare with all teams

View

Metric

Figures cover GitHub contributors during the hackathon window. A co-authored commit counts in full for each author, so per-member totals add up to more than the whole-team figures.

Technology

Found in codeClaimed only
  • AnthropicIn code
  • CSSIn code
  • ExpressIn code
  • HTMLIn code
  • JavaScriptIn code
  • MongoDBIn code
  • PythonIn code
  • ReactIn code
  • Node.jsClaimed

8 of 9 appear in the indexed code. 1 claimed on Devpost could not be matched to code, which may simply mean the tool leaves no trace in the repository.

AI coding agents

No AI coding agent signals were found in this repository.

Detected from committed agent config files and commit authorship. Absence of a signal is not proof an agent was unused.

Codebase size

Source size

52 KB

Source files

20

Counts recognized source files only; vendored directories, binaries and lockfiles are excluded, so this is smaller than the repository on disk.

0 stars