Quick update on the security agents repo I shared last week- just published an 11-minute walkthrough showing exactly how to use them. (For anyone who's been staring at the GitHub repo wondering "okay but how do I actually run these?") The agents check for: • Authentication & authorization gaps • SQL injection & XSS vulnerabilities • Exposed secrets in client code • Missing rate limiting (caught this in the demo) • Database security (RLS coverage) • Dependency vulnerabilities • And about 14 other attack vectors What made me build these agents? We had 3 engineers review a feature last month. Everyone approved. The security agents caught a critical authorization flaw in 6 minutes. It's not about skill. Developers review features for: → Does it work? → Is the code clean? → Do tests pass? Not for: → Can someone bypass this? → What if they modify the request? → Is authorization actually enforced? That's the difference between coding and security thinking. No matter how comprehensive the security checklist is, the tiniest of things might slip. Now we run these agents before we ship anything. Full walkthrough: https://youtu.be/n8IsaWYttp0 Try these on your codebase. Would love to hear your feedback in comments!

Check them out here https://github.com/ZubairLK/security-agents Video on this dropping next week but thought I'd share with the community ahead of time. How are you guys securing your apps? Thanks

Episode 4 of the Bubble to AI code series is now out - authentication in Next.js + Supabase. It is simple once you understand the basics: "Log the user in" in Bubble = SignInWithPassword in code. That's it. Implemented the full authentication flow in 12 minutes: • Login page with email/password fields • Authentication workflow (SignInWithPassword) • Protected dashboard page (server-side redirects) • Logout button and workflow Remember those 302 redirects we set up in Bubble? The "on page load → if current user is logged out → go to login page" workflow? Exact same pattern in code. Same security. Same redirect. Just written differently. What really clicked for me: Authentication isn't complicated. It's the same mental model we already use. Login button → authenticate → redirect to dashboard. The only difference is typing SignInWithPassword instead of clicking "Log the user in." Quick question: How many of you have accidentally left a Bubble page unprotected before? (I definitely have 😅)

Hi all, Just thought I'd start some general discussion on what people are up to these days. I'll go first. - I have recorded a bunch of videos. A marketplace build series for bubble devs switching to AI code. Supabase 101 for Bubble devs etc. But need to prepare a whole bunch more marketing materials around it. - I've been really sick the whole of last week (bad case of food poisoning). Recovering from that & catching up with various projects at the agency What are you all up to & what are you planning on getting done this week? Thanks Zubair

Claude Code builds a feature in 3 minutes. We spend 30 minutes manually clicking around and copy/pasting console logs back to it. That's where Playwright MCP Server comes to rescue. It tests, clicks, and debugs itself. It verifies full feature flows before we even look at it. That’s how you get to production-ready AI code. 10x productivity unlocked!! Check out a quick 5 minutes video where I show my exact setup.