---
name: Find It Later
category: productivity
works_with: [claude, chatgpt, gemini]
difficulty: plug-and-play
version: 1.0
source: original (built for this library; drawn from real drive folders that had collapsed under their own size)
tested: no
---

# Find It Later

## What it does
Turns a folder you've stopped being able to use — a Drive, a Downloads directory, a photo dump, a course library — into something you can actually find things in: a naming pattern that survives a phone screen, a plan for the duplicates, a shallow structure, and one index file. Not a filing system you'll abandon in a week.

## When to use it
- You have hundreds of folders and you search instead of browsing, because browsing stopped working
- Names like "Notes (1)", "Notes final", "Notes final v2" have taken over
- You know the file is in there somewhere and it's faster to remake it

## The skill
```
Act as a librarian organising a collection someone stopped
maintaining. Rules: design for the person I actually am — if a
system needs constant upkeep, I won't do it, so keep it lazy-proof;
never suggest deleting anything without telling me how to check
it's safe first; work only from the folder list I give you and say
what you can't tell from names alone.

My folder/file list (paste it, or describe the mess honestly):
[PASTE NAMES — a screenshot list, a directory listing, or just
describe the categories and roughly how many]
Where it lives: [DRIVE / DROPBOX / LAPTOP / PHONE]
How I mostly open it: [PHONE / DESKTOP / BOTH]
What I come here looking for, realistically: [THE 2-3 THINGS]

Give me:
1. THE NAMING PATTERN: a rule for naming folders and files that
   (a) puts the distinguishing word FIRST, because phone and
   sidebar views truncate the middle and end — "Kids reading —
   grade 2" and "Kids reading — grade 3" are indistinguishable
   when both display as "Kids read...", (b) sorts usefully by
   default, and (c) I can follow half-asleep. Show me my own worst
   examples renamed.
2. THE DUPLICATE PLAN: from my list, the pairs and near-pairs
   ("X" and "X (1)", "final" and "final v2"), how to decide which
   one survives — check dates and sizes first, keep the fuller
   one, and how to park the losers somewhere reversible instead
   of deleting outright.
3. THE JUNK PASS: export artefacts and machine suffixes
   (long timestamp strings, "-001", download numbering) that carry
   no meaning — a single rule for stripping them.
4. THE SHAPE: a flat-as-possible structure — ideally one level of
   real categories, with a small number of pinned favourites at
   the top (a leading character like "_" or "1." forces them to
   sort first). Tell me honestly which of my current categories
   are duplicates of each other under different words.
5. THE INDEX: one plain text or doc file living at the top level,
   listing what each category holds in a line — the thing that
   makes a big collection usable by anyone including future me.
6. THE 20-MINUTE VERSION: if I only do one pass, which of the
   above delivers the most findability for the least work.
```

## Example output
[TO FILL AFTER TESTING]

## Tweaks
- Front-loading the distinguishing word is the highest-value change on a phone; middles and endings get eaten by truncation
- Move duplicates to a "_review" folder rather than deleting — reversible beats tidy, and you'll get braver as nothing breaks
- The index file is what turns a hoard into a library; without it, only you know what's in there, and eventually not even you
- Do the 20-minute version now and the rest never, rather than planning the perfect system and doing none of it