CLAUDE LABJP
FORK — Claude Code 2.1.212 changes what /fork does: it copies your conversation into a new background session with its own row in claude agents, so you can keep working. The old in-session subagent is now /subtaskLIMITS — WebSearch calls are now capped at 200 per session by default, and subagent spawns get the same 200 ceiling, so a runaway search or delegation loop stops on its ownMCPBG — MCP tool calls running past two minutes now move to the background automatically, keeping the session usable. Tune the threshold with CLAUDE_CODE_MCP_AUTO_BACKGROUND_MSPLANFIX — Fixed plan mode auto-running file-modifying Bash commands such as touch and rm without a permission prompt or an SDK canUseTool callbackSONNET5 — Claude Sonnet 5 is running on introductory pricing of $2 per million input tokens and $10 per million output. After August 31 it moves to $3 and $15IPO — Bankers are reportedly lining up investor meetings for Anthropic ahead of a possible public listing as soon as OctoberFORK — Claude Code 2.1.212 changes what /fork does: it copies your conversation into a new background session with its own row in claude agents, so you can keep working. The old in-session subagent is now /subtaskLIMITS — WebSearch calls are now capped at 200 per session by default, and subagent spawns get the same 200 ceiling, so a runaway search or delegation loop stops on its ownMCPBG — MCP tool calls running past two minutes now move to the background automatically, keeping the session usable. Tune the threshold with CLAUDE_CODE_MCP_AUTO_BACKGROUND_MSPLANFIX — Fixed plan mode auto-running file-modifying Bash commands such as touch and rm without a permission prompt or an SDK canUseTool callbackSONNET5 — Claude Sonnet 5 is running on introductory pricing of $2 per million input tokens and $10 per million output. After August 31 it moves to $3 and $15IPO — Bankers are reportedly lining up investor meetings for Anthropic ahead of a possible public listing as soon as October
Articles/Cowork
Cowork/2026-04-17Intermediate

Auto-Organizing Dropbox with Cowork — File Classification, Duplicate Removal, and Naming Conventions

When your Dropbox starts feeling chaotic, let Cowork handle it. Learn how to build an automated system for file classification, duplicate detection, and bulk renaming — and schedule it to run weekly.

Cowork33Dropbox2file organizationautomation95scheduled tasks8file management3

One day you glance at your Dropbox and realize there are 300 PNG files dumped in a Screenshots folder, a Downloads directory full of ZIPs you can't even remember unpacking, and reference images mixed together with app assets from three different projects.

If you're juggling app development and creative work — as I do — this kind of folder chaos builds up faster than you can keep up with manually. Every few months I'd spend an hour sorting things by hand, only to end up in the same state a month later.

Using Cowork changed that cycle entirely. Here's the workflow I actually use to keep Dropbox organized automatically, along with the exact prompts that make it work.

What Cowork Can Do with Dropbox

Before getting into the specifics, it's worth clarifying what Cowork can and can't do with your Dropbox. On macOS, Dropbox is typically mounted at ~/Library/CloudStorage/Dropbox/. Once you've selected that folder in Cowork's workspace settings, Cowork can:

  • Read, move, copy, and delete files across the entire folder tree
  • Identify file types from extensions and metadata
  • Run bash commands for hash-based duplicate detection
  • Apply pattern-based bulk renaming across thousands of files at once
  • Schedule all of this to run automatically on a recurring basis

One thing to note: Cowork operates on the local Dropbox folder directly — it doesn't use the Dropbox API. This means changes are reflected to Dropbox through the normal sync process. Large moves may take a few minutes to propagate to other devices, which is worth keeping in mind if you're organizing while Dropbox is still syncing.

Step 1: Sort Downloads by File Type and Date

The easiest place to start is the most chaotic part of most Dropboxes — the Downloads folder or the root directory itself. Here's the prompt I use:

Please organize files in Dropbox/Downloads using these rules:

- .png .jpg .webp .heic .gif → move to Dropbox/Photos/{YYYY-MM}/
- .pdf .docx .xlsx .pptx .pages .numbers .key → move to Dropbox/Documents/{YYYY-MM}/
- .zip .tar.gz .dmg .pkg → move to Dropbox/Archives/
  (if a folder with the same name already exists nearby, flag the ZIP as likely extracted)
- .mp4 .mov .m4v .avi → move to Dropbox/Videos/{YYYY-MM}/
- Anything else → move to Dropbox/Unsorted/

Before moving anything, show me a preview: total file count per category and example filenames.
Only proceed with actual moves after I confirm.

The preview request is the most important part. Cowork will show you exactly what it plans to do before touching anything, so you catch edge cases — like a .zip file that's actually a deliverable for a client, not an archived download you meant to delete.

After confirmation, the structure ends up looking like this:

Dropbox/
├── Photos/
│   ├── 2026-02/
│   ├── 2026-03/
│   └── 2026-04/
├── Documents/
│   ├── 2026-03/
│   └── 2026-04/
├── Videos/
│   └── 2026-04/
├── Archives/
└── Unsorted/

The month-based subfolders are created automatically based on each file's modification date. If you've been dumping files into Downloads for two years, this single pass can turn hundreds of unsorted files into a browsable, date-indexed archive.

Step 2: Detect and Remove Duplicate Files

Duplicate files are the silent space hogs of any long-running Dropbox account. They're files with different names but identical content — logo_final.png, logo_final_copy.png, logo_USE_THIS_ONE.png — and they're nearly impossible to spot manually at scale.

Cowork can detect them using hash-based comparison via bash:

Please scan Dropbox/Photos for duplicate files.
For any files with identical content, list their filenames, sizes, creation dates, and full paths.
Do not delete or move anything yet — just generate the report.

Under the hood, Cowork runs something like this:

# MD5-based duplicate detection
find ~/Library/CloudStorage/Dropbox/Photos -type f \( \
  -name "*.png" -o -name "*.jpg" -o -name "*.webp" \) | \
  xargs md5 2>/dev/null | \
  awk '{print $4, $1}' | \
  sort -k2 | \
  awk 'prev==$2{print prev_line; print $0} {prev=$2; prev_line=$0}'

Once you have the duplicate report, you can follow up with specific deletion rules:

From the duplicate list, please move the older copy of each pair to Dropbox/DuplicatesToReview/.
Keep the file with the more recent modification date.

Or, if you trust the detection:

Delete the older copy of each duplicate pair. Move to Trash, not permanent delete.

I consistently find between 150–300 MB of duplicates in my Photos folder over a 6-month period. Running this quarterly makes a real difference in Dropbox storage usage — especially if you're on a limited storage plan.

Step 3: Apply Consistent Naming Conventions

Inconsistent file naming is a subtle but real productivity drain. When you're looking for an app screenshot three months after you took it, IMG_5482.png tells you nothing. WallpaperApp_mainscreen_20260310.png tells you everything.

Here's how to enforce naming conventions in bulk:

Please rename all PNG files in Dropbox/AppAssets/Screenshots to this format:
{AppName}_{ScreenDescription}_{DateYYYYMMDD}.png

Use "WallpaperApp" as the app name.
Infer the screen description from the filename if possible (e.g. "home", "settings", "onboarding").
If the original name gives no context, use "screen" as the description.

Show me a before/after table of all filename changes before applying them.

Cowork will analyze each filename, make reasonable inferences, and present a mapping table for you to review. For App Store submission screenshots — where you need separate sets for 6.7-inch iPhone, 12.9-inch iPad, and Mac — this kind of systematic naming makes organizing submissions significantly faster.

A practical tip: run the rename on a single folder first to calibrate the naming output before applying it across your whole asset library.

Running the Whole Workflow on a Weekly Schedule

Once you've tested each step manually and are happy with the results, the obvious next move is automation. In Cowork's Schedule tab, create a new weekly task:

Every Monday at 7:00 AM, run the following Dropbox cleanup:

1. Sort new files in Dropbox/Downloads into Photos, Documents, Videos, and Archives
   folders by type and date. Skip files modified in the last 24 hours.

2. Scan Dropbox/Photos for duplicate files. Save a report to:
   Dropbox/Reports/DuplicateReport_{YYYY-MM-DD}.txt

3. After both tasks complete, display a brief summary:
   - Files sorted (count by category)
   - Duplicates found (count and estimated space)
   - Any files that couldn't be categorized (moved to Unsorted)

Run all steps automatically without confirmation prompts.

The "skip files modified in the last 24 hours" clause is a practical safeguard — it avoids moving files you're actively working on. Monday morning is a natural time because Dropbox has had the weekend to fully sync any changes from your other devices.

For more on setting up scheduled tasks, the Cowork Scheduled Tasks Guide walks through the full configuration process.

Building a Folder Report Before You Start

Before reorganizing a folder you've been using for years, it's helpful to get a picture of what's actually in there. Cowork can generate a quick inventory:

Please scan Dropbox/ (top 2 levels only) and give me a summary:
- Total file count per top-level folder
- Largest 10 files by size
- File types that appear more than 20 times
- Folders that haven't been modified in over a year

Don't move or change anything — just the report.

This report is useful in two ways. First, it helps you decide where to focus your cleanup effort. If Dropbox/Old Projects/ has 8 GB of files untouched since 2023, that might be a better starting point than Downloads. Second, it surfaces files you didn't know were taking up space — large video exports, uncompressed PSDs, old disk images.

Running this audit before a bulk sort prevents the frustration of carefully organizing a folder, only to realize the real clutter was somewhere else entirely.

Common Pitfalls to Avoid

Always move to a review folder instead of deleting outright. When you're first building these workflows, replacing "delete" with "move to Dropbox/Trash/" in your prompts adds a safety net. Cowork is accurate, but your own naming patterns might have edge cases you didn't anticipate. Give yourself a week to review the moved files before making the deletion permanent.

Watch out for Dropbox selective sync. If you have selective sync enabled, some folders on your local machine may not be downloaded. Cowork won't be able to see files that haven't been synced locally. Before running a large-scale organization, make sure the relevant folders are marked for local availability in Dropbox Settings.

Test on a small folder first. Before applying any bulk rename or move to your entire Dropbox, pick a folder with 20–30 representative files and run the prompt there. This is especially important for rename operations — you want to make sure the naming inference is behaving the way you expect.

Going Further

The file operations used here build on the basics covered in File Management and Desktop Operations. For more on combining file workflows with recurring schedules — including error handling and logging — Automating Tasks with Cowork has additional examples that apply to any folder, not just Dropbox.

If this is your first time setting up a Cowork file workflow, start with just the Downloads folder sort. A single clean pass is usually enough to understand how the system works — and to realize how much time you've been spending on something that can now just happen on its own.

Share

Thank You for Reading

Claude Lab is ad-free, supported entirely by members like you. We publish practical guides daily with implementation code, benchmarks, and production-ready patterns. If you've found it useful, we'd love to have you on board.

  • Copy-paste ready implementation code
  • New advanced guides published daily
  • $5/mo or $10 for lifetime access
View Membership →

If you found this article helpful, a small tip ($1.50) would mean a lot to us. Your support helps keep this site ad-free and covers server and hosting costs.

Related Articles

Cowork2026-06-21
Why Cowork's bash Says 'No Such File' When Finder Shows It Right There
Connect a cloud-synced folder to Cowork and bash sees empty placeholders while cat fails. Here is how on-demand materialization actually works, and the design patterns that keep your automations from silently dropping data.
Cowork2026-06-23
Stopping an Unattended Writer From Publishing the Same Article Twice
When a Cowork scheduled task generates articles every day, the real danger isn't a crash — it's quietly publishing a piece that overlaps with one from a few days ago. Here is a gate that compares slug similarity and the day's log before publishing, built from a near-miss I caught this morning.
Cowork2026-03-20
Running 4 Sites with 600+ Articles on Autopilot Using Cowork — A Solo Developer's Real Automation Story
How I use Claude's Cowork mode to automatically generate and publish articles across 4 AI knowledge base sites. Covers scheduled task design, skill files, real failures, and lessons learned.
📚RECOMMENDED BOOKS
Build a Large Language Model (From Scratch)
Sebastian Raschka
LLM Dev
Prompt Engineering for LLMs
Berryman & Ziegler
Prompting
AI Engineering
Chip Huyen
AI Eng
* Contains affiliate links
See all →