I would kill to see pandoc integration into Scrivener, and by “kill” I mean: coerce everyone I know into buying a copy.
(I’d also be willing to help code that integration.)
… and I’d willing to write the documentation.
+1
The last time I checked, Scrivener’s MMD output is already Pandoc-markdown compatible. Here is what it produces as syntax:
-
Headings, via the Formatting pane: compatible (both use Atx style).
Advanced header syntax can be implemented with with prefix/suffix usage; same as with MultiMarkdown users. Scrivener supports placing these both inside and outside the hash marks. So if you wanted an unnumbered heading below a certain level, you could apply a prefix of {-} outside of the hash. -
Images (with optional captions): compatible, though avoid resizing images in the editor as that will cause Scrivener to use MMD’s image size attribute syntax, which Pandoc does not support. Supposedly the link_attributes extension will resolve that, but in my testing it does not correctly handle MMD style image size attributes. To be extra safe, you could write a Replacement regex that strips out the width and height values, but like I say that shouldn’t necessary unless you mess with image sizes in the editor.
-
Footnotes: compatible.
-
Meta-data: compatible if you use the mmd_title_block format extension in Pandoc, and if you’d rather stick with Pandoc meta-data blocks, you needn’t use the Meta-Data panel in the compiler—just add a document called “Meta-Data” at the very top of your compile group and type it in to that file. Scrivener will treat that file “As-Is”, and make sure to keep it spaced out from the main content with an empty line.
-
Verbatim: compatible. Paragraphs enclosed in a Preserve Formatting field will have one tab inserted in front of every such paragraph. Ranges of text within paragraphs will be enclosed in backticks.
-
Section breaks: compatible, as used in the Separators pane. Will insert a series of hyphens.
As for everything else: even MultiMarkdown users have to type that stuff in, you’d just occasionally be typing it in differently. 
And of course, at its most basic, using “Plain-text” as the output, you can use Scrivener to write anything from a C++ header file to an XML file to a Pandoc document. Then you have full control, but you don’t get any of the conveniences from the above. Do note you can even use various compile features to generate syntax for you. Replacements and the Formatting pane are quite powerful.