Safe automation script for a repetitive task

You are a careful developer who writes automation scripts that are safe to run and easy to maintain. Goal: a working script for the task below, in the language I name. Inputs: - Task to automate, described step by step: {{task_description}} - Language and runtime version: {{language_and_runtime}} - Input files, folders or data sources, with a sample: {{inputs}} - Expected result: {{expected_result}} - Safety limits (what must never be deleted, overwritten or sent): {{safety_limits}} Steps: 1. Restate the task and list assumptions about the files, formats and permissions. Ask at most 3 questions only if a missing detail could cause data loss. 2. Design before coding: list the steps, the failure points (missing file, bad format, network error, rate limits) and how each will be handled. 3. Write the script with clear names, comments only where the reason is not obvious, settings at the top, and no secrets in the code (read them from environment variables). 4. Include a dry-run mode that prints what would happen without changing anything, make it the default, and log every action with a summary at the end. 5. Never delete or overwrite without a backup or explicit flag. Respect the terms of any site or service the script talks to and keep request rates modest. 6. Walk through a test using my sample input and show the expected output. Output format: - Plan (short list) - The complete script in one code block - How to run it, including a dry run and a real run - What the script does not handle Quality bar: simple and readable over clever; no unneeded dependencies; tell me which parts you could not test.

You are a careful developer who writes automation scripts that are safe to run and easy to maintain. Goal: a working script for the task below, in the language I name. Inputs: - Task to automate, described step by step: {{task_description}} - Language and runtime version: {{language_and_runtime}} - Input files, folders or data sources, with a sample: {{inputs}} - Expected result: {{expected_result}} - Safety limits (what must never be deleted, overwritten or sent): {{safety_limits}} Steps: 1. Restate the task and list assumptions about the files, formats and permissions. Ask at most 3 questions only if a missing detail could cause data loss. 2. Design before coding: list the steps, the failure points (missing file, bad format, network error, rate limits) and how each will be handled. 3. Write the script with clear names, comments only where the reason is not obvious, settings at the top, and no secrets in the code (read them from environment variables). 4. Include a dry-run mode that prints what would happen without changing anything, make it the default, and log every action with a summary at the end. 5. Never delete or overwrite without a backup or explicit flag. Respect the terms of any site or service the script talks to and keep request rates modest. 6. Walk through a test using my sample input and show the expected output. Output format: - Plan (short list) - The complete script in one code block - How to run it, including a dry run and a real run - What the script does not handle Quality bar: simple and readable over clever; no unneeded dependencies; tell me which parts you could not test.

Plan: scan the downloads folder, group files by extension, move them into subfolders, skip files modified in the last hour, log each move. Script: (complete code block with settings at top, dry-run enabled by default, a log line per file, and a final count) How to run: 1. Preview: run the script with no flags; it only prints planned moves. 2. Real run: add the --apply flag. Not handled: files in use by another program (they are skipped and logged); duplicate names (a number is added to the new name). Untested: behaviour on network drives.