Reviewing Claude Code output before pushing directly to prod

The 2-second LGTM before pushing to prod is peak engineering. At least now you can blame the AI.

"I mean, it compiles?"

Thank you for todays laugh.

🫡

As long as you said "make no mistake" means its fine right? 🙄

Reviewing? What’s that, sounds ghetto.

just checking if it exposed creds

.env in .gitignore, ez pz

@boris0chuck im sorry

Just let Claude ssh to the prod env already...don't be the bottleneck of the system 👾

Before?

This is a pretty good frisking technique if you’re looking for a ballistic missile launcher.

"Reviewing" = scrolling past the diff and hitting merge. We all do it lol

Any reason why he acted slowly to the last man 😂

Safe > sorry

The scariest part isn't that people push AI code without reviewing. It's that most devs were already pushing code they barely understood before AI existed. Claude just made the skill issue visible.

Kind of funny and scary at the same time.

Wait, you’re still reviewing the code?

🤣

That’s actually straight to the point thanks mate this is a good one

hair loss will be cured with claude

when my app gets completely hacked, i will just copy and paste the error message into claude

Spot on, this is how it’s done

Meta security measures

Seems reasonable…what could possibly go wrong

Ngl, this guy probably can read body language as a giveaway vs anything

banger

ship now, cry later

@MetalLegBob got a great tool that will help a bunch here. Sanity and security check against Multiple metrics!

Wait.. you guys review the code before pushing it?? I thought we only review the code when it brake in the production!

Hate to say it but opus 4.6 sucks!

Nah dawg, I don’t trust Claude like that. This is me with codex

Why? Just use the /make-no-mistakes skill

@stunimmo

😂😂😂😂 @SourceD20

csnt really review it if you cant read code ahaha

high trust society type beat