Write a shell script to automate a repetitive task

Write a shell script that automates the task below for {{user}}, who will run it {{how_often}}, so that it gets done the same way every time without anyone retyping commands.

The steps I do by hand today:
"""
{{steps}}
"""

Operating system and shell: {{shell}}
Tools available: {{tools}}
Where it runs: {{where_it_runs}}

Example input (file names or a folder listing):
"""
{{example_input}}
"""

The script must:
1. Stop at the first error: start with set -euo pipefail in bash or zsh, or the equivalent for this shell.
2. Take paths and options as arguments or as variables at the top, never buried in the middle, and print a usage line when a required one is missing.
3. Offer a dry-run option that prints each action without doing it.
4. Quote every variable, so names with spaces work.
5. Be safe to run twice: a second run must not duplicate, overwrite or delete what the first run did.
6. Check that a target is what it should be before deleting or overwriting it, and never run rm -rf on a path built from a variable that could be empty.
7. Print one short line per step and a summary at the end, and exit with a non-zero status on failure.

Return the script in one code block with a comment above each section, then the exact commands to make it executable, run a dry run and run it for real, then up to 5 assumptions you made.

A good script is one I can read in two minutes and trust to run unattended. For example, the form of one log line: "Copied 42 files to /backups/<date> (skipped 3 already present)."

Use only commands available on {{shell}} plus the tools listed. The macOS and Linux versions of sed, date and find take different options, so write for the system named and say which one you wrote for. Do not use sudo unless a step needs it, and explain why if it does. If a step is ambiguous, such as what to do when a file already exists, write the safest choice, mark it with a # CHECK comment and list it with your assumptions.