AI writes code that works.

@itsthatlady.dev

29.0K followers · usually 1.5K views a post

Views

50.5K

Likes

5.2K

Comments

92

Shares

990

Selling · SaaS

Replitseen in 54 videos →https://replit.com/refer/itsthatladydev ↗

Why it worked

The video provides high-value, actionable technical advice that addresses a common pain point for developers using AI tools, positioning the creator as an expert while naturally integrating the platform used for the demonstration.

Summary

The creator explains three essential security checks for developers using AI to write code, emphasizing the importance of environment variables, established authentication providers, and using AI to audit code for vulnerabilities. The video concludes by advising viewers to save these tips for future development projects.

Structure

1Identify the problem of AI-generated code security
2Explain the risk of hardcoding API keys
3Recommend using battle-tested authentication providers
4Demonstrate how to use AI to audit code for vulnerabilities
5Final advice on treating security as a standard practice

Product placement

The creator uses Replit throughout the video to demonstrate code editing and AI prompting.

Original caption

AI writes code that works. It doesn't write code that's safe 🚨 Here are 3 security checks every vibe coder needs to run before shipping - each takes 2 minutes, skip them and you're handing hackers the keys to your app. Save this. You'll need it at 2am when something breaks 😅 #VibeCoding #WebSecurity #AITools #techtok #CodingTips

More from @itsthatlady.dev

Their other posts in the index, biggest breakout first.

Call to action

Save this for the next build.

On-screen text

AI writes code that works It doesn't write code that's safe 1 2 3 Skip them and you're handing hackers the keys When you paste API keys into your chat, you are potentially exposing them 1. Use environment variables Tell your AI to use processes.end Never hardcode credentials One leaked key and the hacker owns your entire app AI will build you a custom login system that looks like it works, but has holes you can't see 2. Use battle tested providers Supabase Auth, Clerk, NextAuth Never let AI invent authentication from scratch Auth is where 90% of data breaches start Open your code and prompt your hacker You're a hacker trying to break this app, tell me every vulnerability and how to exploit it, then tell me the fix AI knows attack patterns 3. Start using it for breaking as well Security isn't paranoia It's showing respect for people's data These aren't advanced techniques These are the basics that most AI built apps are missing right now Save this for the next build, you'll need it at 2 a.m. when something breaks

Transcript

AI writes code that works. It doesn't write code that's safe. I check these three things in every single project. Each takes two minutes. Skip them, and you're handing hackers the keys to your app. When you paste API keys into your chat, you are potentially exposing them. 1. Use environment variables. Tell your AI to use processes.end. Never hardcode credentials. One leaked key and the hacker owns your entire app. AI will build you a custom login system that looks like it works, but has holes you can't see. Instead, 2. Use battle tested providers. Supabase Auth, Clerk, NextAuth, never let AI invent authentication from scratch. Auth is where 90% of data breaches start. Now this is the check most vibe coders never think to do. Open your code and prompt your hacker: You're a hacker trying to break this app, tell me every vulnerability and how to exploit it, then tell me the fix. AI knows attack patterns. Most people use it for building, start using it for breaking as well. Security isn't paranoia. It's showing respect for people's data. These aren't advanced techniques. These are the basics that most AI built apps are missing right now. Save this for the next build, you'll need it at 2 a.m. when something breaks.