Audit Time!
Lets face it: we live in a world of AI and there's no turning back. For the last year or two, I've avoided AI for anything serious. I don't use it at work, I enjoy learning and the thrill of struggling with a new thing as it feels so much better when things fall into place afterwards. But between my professional life and this project, I've learned there is simply too much I don't know yet. While I continue to learn, I've realized I also must progress at at rate that keeps up with demand.
This is where Claude comes in for me. I've played with it for a while, testing how well it can answer my questions where I provide limited information about Anitrac and zero insight into the code. I have to admit: this is awesome. I've been able to learn so much and retain it easier as I can tailor how things are explained or ask follow up questions immediately. That's when I decided to get Anthropic's cheapest subscription to help me learn what I needed for Anitrac. My usage for AI for Anitrac splits into two factors: planning (potentially more on this in a future blog update) and audits.
Whats an Audit...
...and not in a literal sense. Being in this industry, you've more than likely had your fair share of them, but what does that mean for Anitrac? I don't like AI change my code-base, I like to take pride in what I am creating. Test can only get you so far though. And while all 55 of my current test passed with the last update of the theme, there's plenty of blind holes I missed as a solo-dev. Starting this year, I zip my project and have Claude check for things I overlook and create a markdown file of all issues so I can tackle them. This is done every other month so I can not only catch issues early on before the scope of the project increases to far past the issue, but also to ensure Anitrac will be delivered at it's best, not just "good enough to work".
Audits of Tomorrow
While Anitrac is in development and in alpha/pre-beta mode, these audits will remain internal. But after an full beta is reached, this will be reformatted as not to give too much away about the code base for now and posted along with a change log for transparency. And for the eagle eye viewers, you may notice something about the test: versioning. The backend will be an open API for specific users and I feel sharing vulnerabilities from this audits with that user base for their integrations.
