Romaji to Kana Converter: Astro — Chapter 8, Exercise 1
=====================================================================

TASK
Write your own complete my-tools/romaji-converter.yaml spec file
matching MT3's schema, including a real example block, based entirely
on what this course actually built in Chapters 2 through 6.

SOLUTION

name: romaji-converter
description: |
  A single-page tool that converts typed romaji (Latin-alphabet Japanese)
  into hiragana and katakana, live as the user types. Handles the real
  gojuon grid, voiced/semi-voiced sounds, yoon combinations, long vowels,
  sokuon (small tsu), and the ha/he/wo particle exceptions. Client-side
  only -- a plain input box, and two output areas for hiragana and
  katakana, styled to match the site's own established tool palette.
  Kanji conversion is explicitly out of scope.
example:
  input: konnichiwa
  hiragana_output: konnichiha (kana characters)
  katakana_output: konnichiha (katakana form)

Every field in this spec traces back to a real, already-built chapter
rather than an invented feature: the description's own "real gojuon
grid... voiced/semi-voiced sounds, yoon combinations" comes directly
from Chapter 2's KANA_MAP, "long vowels, sokuon... particle exceptions"
from Chapter 3, "single-page... live as the user types" from Chapter 4,
and "client-side only" from Chapter 5's own real, measured verdict.
The example reuses "konnichiwa" specifically because Chapter 3's own
lexicalized-exception dictionary is what makes this exact word convert
correctly in the first place -- a genuinely representative example, not
an arbitrary pick.

WHY THIS WORKS AS AN ANSWER
----------------------------
It follows MT3's real required schema (name, description, example) and
grounds every claim in the spec's own description directly in a specific
chapter of this course rather than describing hypothetical functionality
-- confirming the spec is a faithful summary of what was actually built,
not a wish list.