feat(code): add Kotlin comment extraction

[anandghegde]

Closes #1125.

Adds Kotlin (.kt, .kts) to the tree-sitter comment extractors, following the same layout as the Elixir support in #1151. It extracts // line comments and /* */ block comments, KDoc /** */ included (the grammar parses KDoc as an ordinary multiline_comment). The grammar is already vendored in go-tree-sitter, so there's no new dependency.

Two small things beyond the new language:

• The padding lists /** as a marker. With plain cStyle, an alert in a one-line KDoc comment (/** ... */) lands two columns short. JavaScript has the same problem today (/** foo */ in a .js file), which I've left alone here.
• TestComments now pairs in/N.ext with out/N.json by file name instead of by position. 10.kt sorts before 2.py, so the positional pairing broke as soon as there were more than ten fixtures.

Tests:

• testdata/comments/{in,out}/10.kt, 11.kts: line-comment runs, a KDoc block with a tag, an inline block comment, a trailing comment, a one-line KDoc, and strings containing TODO and // that must not be extracted.
• testdata/e2e/lint.yaml: a kotlin case against testdata/fixtures/formats/test.kt, with all eight columns checked.

make test passes locally (the converter-dependent cases were skipped). I also ran a built vale with a one-word existence rule on a .kt and a .kts file: it flagged the word only in line, block and KDoc comments, not in identifiers or strings, and the line/column numbers were right.

Assisted-by: Claude Code

[github-actions]

The license agreement is signed.
_{Posted by the CLA Assistant Lite bot.}

[github-actions]

Before this is reviewed, please read the AI-Assisted Contributions section of the contributing guide and make sure this change meets it: you can explain every line, the description is your own and brief, it fixes a bug you hit or a feature agreed in an issue, it does not touch undocumented options, and any substantial tool use is noted.

[anandghegde]

I have read the CLA and I agree to it