SuperAPI AI Directory

i18n-pseudo-workflow

Office & productivity

View repository

Updated 10-15 22:15

About this skill

i18n-pseudo: $ARGUMENTS

Pseudo-translate the given locale file(s) with the i18n-pseudo CLI, wire the result into the app as a test locale, and report what the test exposes.

Step 1 — Setup check

command -v i18n-pseudo || npm install -g @i18n-agent/i18n-pseudo

Step 2 — Pick the preset (if the user didn't)

Situation Preset
Quick scan for untranslated/hardcoded strings default
Target languages with long words (de, fi) / overflow worries layout
CJK targets or font/encoding concerns charset
Arabic/Hebrew support rtl
Pre-release full sweep full

Step 3 — Generate

i18n-pseudo <files> --preset <preset> -o <output-dir>

Default output dir: pseudo/ next to the input. Never use --in-place on the real reference locale unless the user explicitly asks; if they do, keep the .bak files.

Step 4 — Wire it up and inspect

  • Register the pseudo output as a test locale in the project's i18n config (e.g. a qps or en-XA locale) so the app can actually render it.
  • Verify placeholders survived: grep the pseudo output for the placeholder patterns present in the source ({, {{, %s, ${); they must be intact.
  • If the app can be run/screenshotted, render key screens in the pseudo locale: any text WITHOUT brackets/accents is a hardcoded string that bypassed the i18n layer.

Step 5 — Fix or report

List every hardcoded string found with file:line. If the user asked for fixes, wrap them (use /i18n-wrap for bulk wrapping); otherwise report the list and the recommended next step.