Build a File Organizer Agent

Objective

Build a Python agent that uses Claude's Computer Use capability to organize files in a messy Downloads folder by moving them into categorized subfolders.

Background

This lab combines visual GUI automation with bash commands for an efficient hybrid approach. The agent should use bash for bulk file operations while using the computer tool for visual verification.

Requirements

Create a class FileOrganizerAgent with the following methods:

1. `analyze_folder(path: str) -> dict`

Scans the folder and returns file counts by category:

{
    "images": ["photo1.jpg", "screenshot.png"],
    "documents": ["report.pdf", "notes.txt"],
    "videos": ["clip.mp4"],
    "archives": ["backup.zip"],
    "other": ["random.xyz"]
}

2. `create_category_folders(base_path: str) -> list[str]`

Creates subfolders for each category if they don't exist. Returns list of created folder paths.

3. `organize_files(path: str) -> dict`

Moves files to appropriate category folders. Returns summary:

{
    "moved": 15,
    "skipped": 2,
    "errors": []
}

4. `generate_tool_calls(path: str) -> list[dict]`

Returns the sequence of Computer Use tool calls needed to:

Open file manager at the specified path
Verify the organization visually
Take a screenshot of the result

File Categories

Category	Extensions
images	.jpg, .jpeg, .png, .gif, .bmp, .svg, .webp
documents	.pdf, .doc, .docx, .txt, .md, .xlsx, .pptx
videos	.mp4, .avi, .mkv, .mov, .webm
audio	.mp3, .wav, .flac, .aac, .ogg
archives	.zip, .tar, .gz, .rar, .7z
code	.py, .js, .ts, .html, .css, .json
other	everything else

Expected Tool Call Format

{
    "type": "tool_use",
    "name": "computer",
    "input": {
        "action": "type",
        "text": "nautilus /home/user/Downloads"
    }
}

Hints

Use os.path.splitext() to get file extensions
Use shutil.move() for moving files
The bash tool is faster for file operations than GUI clicks
Use Computer Use for visual verification at the end

Instructions