Claude's biggest flaw is it forgets everything. I fixed that permanently with 3 files.

Now it remembers my preferences, my writing style, my corrections - and it gets smarter every conversation.
Here's the setup (takes 5 minutes):
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Start a brand new desktop folder, and build these three markdown files:
1. Instructions. MD - tells the model how to act
Example:
## Who you are
## What you do
## Rules
## What good outputs look like
## Memory
In this folder, include: "UPDATE Memory . MD with my preferences over time."
This line is crucial; it's how you get Claude to create a running memory log of all your data.
2. Memory. MD - this is the "brain" of Claude, and what gets continuously updated over time.
Example:
## Preferences
## Corrections
## Patterns
Now, whenever you say something like "stop using em dashes" in Claude Code, it will automatically update your Memory. MD file to reflect that preference because your instructions in the first .MD file told it to do so.
You can then take this Memory. MD and use it across any LLM or Claude chat without having to re-explain anything - super useful for big projects or transferring data.
3. Context. MD - any context the model needs based on your project
This is one of the most powerful memory setups you can run inside of Claude.
Make sure to save this so you don't forget it.
(tip: copy & paste into Claude co-work to guide you and help build this for you).