[ 🏠 Home / 📋 About / 📧 Contact / 🏆 WOTM ] [ b ] [ wd / ui / css / resp ] [ seo / serp / loc / tech ] [ sm / cont / conv / ana ] [ case / tool / q / job ]

/css/ - CSS Masters

Advanced styling, animations & modern CSS techniques
Name
Email
Subject
Comment
File
Password (For file deletion.)

File: 1786360302075.jpg (142.1 KB, 1024x1024, img_1786360292957_d8o4o5ww.jpg)ImgOps Exif Google Yandex

6426d No.2006

fr anthropic is making auto mode the default bc apparently we can't be trusted to supervise the agent. it used to be that u had to manually approve every single change, but now the agent is just gonna run wild. it sounds like a recipe for broken layouts . i wonder if this means we will see more accidental
display: none;
errors in our commits. it feels like we are moving toward a future where the agent just does what it wants w/o any oversight. is anyone else actually comfortable letting an agent commit directly to production?

link: https://thenewstack.io/claude-code-auto-mode/

d18c9 No.2007

File: 1786361940958.jpg (129.56 KB, 1024x1024, img_1786361901616_64bfoc46.jpg)ImgOps Exif Google Yandex

>>2006
the real nightmare isn't just the accidental
display: none;
but when it silently overwrites a complex z-index stack or breaks an existing media query. if you aren't running a strict visual regression testing suite like Percy or Chromatic, you're basically flying blind. i started forcing all agent commits thru a mandatory visual diff check b4 they even hit the staging branch.
>it's not about trusting the code, it's about trusting the test coverage.

**if your tests are trash, auto-mode is just an automated way to break your site faster


[Return] [Go to top] Catalog [Post a Reply]
Delete Post [ ]
[ 🏠 Home / 📋 About / 📧 Contact / 🏆 WOTM ] [ b ] [ wd / ui / css / resp ] [ seo / serp / loc / tech ] [ sm / cont / conv / ana ] [ case / tool / q / job ]
. "http://www.w3.org/TR/html4/strict.dtd">