Quick Start Guide

The quickest way to run your first code audit is through a GitHub Action workflow that will push your repository to CodeAudits.ai, where you can then review audits for your codebase.

name: Push codebase for LLM audits (public)
 
on:
  workflow_dispatch:
 
jobs:
  parse-repository:
    runs-on: ubuntu-latest
    name: Parse this repository for CodeAudits
    steps:
      - name: Checkout
        uses: actions/checkout@v4
 
      - name: Parse Repository
        uses: frogermcs/codeaudits-parse@v1.1.0
        id: parse
        with:
          style: markdown
          compress: false
          push-to-codeaudits: true
      
      - name: Upload Prompt File as Artifact
        uses: actions/upload-artifact@v4
        with:
          name: parsed-repo.txt
          path: parsed-repo.txt

You can find the workflow file in the CodeAudits repository: llm_audit_public.yml

This workflow uses the CodeAudits GitHub Action to parse your codebase and send it to CodeAudits.ai for analysis.

Audit through Google Colab (no changes to your repo)

You can also try CodeAudits through a Google Colab notebook: Code Audits - Quickstart Guide. To test it out, run all the code via Runtime -> Run All.

Example Audit

Here’s the example audits here.

Who Should Use It?

• If you have a public repository on GitHub.
• If you want to check out CodeAudits.ai in under five minutes.
• If you’re okay with using the free-of-charge Google Gemini API (your submitted data may be used to train Google’s models).

Set Up Your Gemini API Key

By default, the app uses a free-to-use API key for the Gemini API. However, it’s likely this key will expire or its limits will be exceeded eventually.

If you want to keep using the Code Audits app, you can set up your own Gemini API key, which will be stored in your browser’s session.

To quickly generate a Gemini API key, visit: https://aistudio.google.com/.