extract.sh

#!/bin/bash
# @parameters file workspace

# Set locale to UTF-8
export LC_ALL=en_US.UTF-8
export LANG=en_US.UTF-8

# Get the current file and workspace paths
current_file="$1"
workspace="$2"

# Get the selected text from the clipboard
selected_text=$(pbpaste)

# Extract the timestamp from the selected text
timestamp=$(echo "$selected_text" | sed -n 's/^## \([0-9]\{10\}\).*/\1/p')

if [ -z "$timestamp" ]; then
    echo "Error: No timestamp found in the selected text"
    exit 1
fi

# Create the new file path
new_file_path="$workspace/_Posteingang/$timestamp.md"

# Extract the content, replace ## with # for the first heading
extracted_content=$(echo "$selected_text" | sed '1s/^## /# /')

# Save the extracted content to the new file with UTF-8 encoding
echo "$extracted_content" | iconv -f UTF-8 -t UTF-8 > "$new_file_path"

# Create the wiki link
first_line=$(echo "$selected_text" | sed -n '2p' | cut -c 1-16)
wiki_link="## $timestamp\n\n[[$timestamp]]"

# Copy the wiki link to clipboard
echo -e "$wiki_link" | iconv -f UTF-8 -t UTF-8 | pbcopy

# Notify the user
echo "Extracted note saved to: $new_file_path"
echo "Wiki link copied to clipboard. Please paste it in place of the selected text."

I had problems with non-ascii code in the initial script.
This one uses UTF-8 explicitely and works with non-ascii (like German Umlaut and emojis) super well.