They seem to do find-and-replace, hence many strange behavior. Macana does not do the same parsing as it’s quite unintuitive and AST-unfriendly.

Basics

> [!info]+ Title
> Body

[!info]+ Title Body

Escapes

> \[!info]+ Title
> Body

[!info]+ Title Body

> [!info]\+ Title
> Body

[!info]+ Title Body

\> [!info]+ Title
> Body

> [!info]+ Title > Body

Edge cases

No whitespace before the marker

>[!info]+ Title
>Body

[!info]+ Title Body

Phrasing semantics continues from title

> [!info]+ **Bold _Italic
> Body, Italic_ Bold**

[!info]+ Bold Italic Body, Italic Bold

Sequential markers

> [!info]+ Title
> [!info]+ Title
> Body

[!info]+ Title [!info]+ Title Body

Whitespace in title

> [!  info  ]+ Title
> Body

[! info ]+ Title Body

Whitespace before exclamation mark

> [ !info]+ Title
> Body

[ !info]+ Title Body

Whitespace before foldable sign

Seriously?

> [!info] + Title
> Body

[!info] + Title Body

My guess is they incorrectly re-parse the rest of the first paragraph as block content, like this:

+ Title\nBody

• Title Body

Block quote works fine.

> + Title
> Body

• Title Body