I think you misunderstood. The suggested format is basically a “plaintext snippet” with YAML frontmatter, i.e. everything after the second ^---$\n
will be considered literal snippet content, and therefore does not require any YAML-specific considerations such as whitespace. This is common practice in some markdown parsers for example.
YAML is only used for the snippet metadata, i.e. tab trigger, name and scope selector. (In fact, SaneSnippets doesn’t even use a YAML parser but simple regular expressions.)
I added a few more examples to the github issue, but here is one that should show this more clearly:
---
description: Enumeration (annotated)
tabTrigger: enum
scope: text.tex.latex
---
\begin{enumerate}${1/(.+)/\[/}$1${1/(.+)/\]/}
\item $0
\end{enumerate}
For comparison, this is the generated .sublime-snippet
:
<snippet>
<description>Enumeration (annotated)</description>
<tabTrigger>enum</tabTrigger>
<scope>text.tex.latex</scope>
<content><![CDATA[
\begin{enumerate}${1/(.+)/\[/}$1${1/(.+)/\]/}
\item $0
\end{enumerate}
]]></content>
</snippet>