Joshua

im just building, thanks

[Joshua @JoshuaThirteen]

Claude Code connected to apps via MCP is like having a full‑stack dev at your fingertips, but most people aren't using it correctly

Its not your fault, either- setting up MCPs in CC is not as easy as an IDE like Cursor/Windsurf

But done correctly, it will 10x your code 🧵

[Joshua @JoshuaThirteen]

@eCom_Amin MCP

[Joshua @JoshuaThirteen]

@blvckledge Demand

[Joshua @JoshuaThirteen]

@eCom_Amin Guide

[Joshua @JoshuaThirteen]

@itsalexvacca Ads

[Joshua @JoshuaThirteen]

@lifemaximised Campaign

[Joshua @JoshuaThirteen]

@codyschneider AEOdashboard

[Joshua @JoshuaThirteen]

@celispj @judgeme Love using Judgeme on my store for years, but realizing it only does Product Ratings

Any plans to incorporate Google Store Ratings as well so we can link to Merchant Center?

[Joshua @JoshuaThirteen]

@austriker27 @Shopify Took them long enough! Can’t wait to test it out

[Joshua @JoshuaThirteen]

@Shopify FINALLY

[Joshua @JoshuaThirteen]

@filippkowalski Missing Nox

[Joshua @JoshuaThirteen]

@mikefutia Skill

[Joshua @JoshuaThirteen]

@mikefutia Ads

[Joshua @JoshuaThirteen]

@austriker27 Need to start it in —dangerously-skip-permissions and run /loop and tell it do not stop until you’ve completed every step, run tests, verified tests, fixed bugs, run /security-review, and test frontend (if applicable) directly using Claude in chrome

Heavy handed but this works

[Joshua @JoshuaThirteen]

This was the first message on Opus 4.6 with high reasoning and ultrathink enabled

Unbelievably unacceptable

[Joshua @JoshuaThirteen]

I've been a huge supporter and user of claude code @claudeai but wtf is this disaster?

[Joshua @JoshuaThirteen]

@austriker27 And you can use the Shopify CLI to directly edit your liquid code files? If so, I need to switch to this method ASAP

[Joshua @JoshuaThirteen]

@austriker27 Shopify CLI in Claude Code?

[Joshua @JoshuaThirteen]

@0x0SojalSec Which “benchmarks” does this 0.9B OSS model beat Gemini? I can assure you that this is no where close to OCR capabilities of major lab models and it’s disingenuous to represent it as such

Just look at OCR Arena- GLM-OCR is literally last place by a huge margin 😭

[Joshua @JoshuaThirteen]

@skalskip92 Def impressive for a small, local model, but disingenuous to claim it beats Gemini or other lab models when the data clearly shows otherwise

Literally bottom of the rankings in OCR arena