How can I stop wp_update_post messing up HTML example code?

I’m showing little pieces of example HTML/PHP code on the frontend (with syntax highlighting). The entries are displayed as FAQ custom posts, and have a plugin (CMS Tree page) to alter the menu order, at which these are displayed. Whenever I change the order, it calls a wp_update_post, updating ID, menu_order, post_parent, post_type. For debug, I output the post_content property right before updating, and right after, and it gets changed in a horrible way.

Before (correct):

</body>

After (messy):
</body>

If I do it a couple of times it’ll become living hell to revert:
</body>

It looks exactly like htmlspecialchars() with ENT_NOQUOTES has been cast on it…
Please note that these are from the “source code” view and not the frontend. The browsers correctly handle the “Before” state, and render it as
</body>

How can I prevent it going messy?

It was https://wordpress.org/plugins/syntaxhighlighter/ plugin’s fault. There are some functions in it that have something to do with this, namely encode_shortcode_contents_slashed_noquickedit, encode_shortcode_contents_callback.

Now I replaced it to https://wordpress.org/plugins/crayon-syntax-highlighter/ but by the time I confirmed it was the other plugin’s fault I had already written a solution.

function wpse190396_insert_post_data($data, $postarr){
    if($postarr['filter'] == 'db'
        && ($data['post_type'] == 'fix' || $data['post_type'] == 'faq')
        && (strpos($data['post_content'],'<') !== false
            || strpos($data['post_content'],' ') !== false
            || strpos($data['post_content'],'>') !== false
            )){
        $data['post_content'] = htmlspecialchars_decode($data['post_content']);
    }
    return $data;
}

add_action( 'wp_insert_post_data', 'wpse190396_insert_post_data', 10, 2 );