Guides

Giving your assistant how you work

Short answers to the questions that come up when an AI assistant keeps forgetting who it is working with.

Claude Code memory: what it remembers, and how to give it the rest
How Claude Code carries context between sessions — CLAUDE.md, its own notes, skills and MCP — where each one runs out, and how to hand it how you actually work.
How to write a CLAUDE.md that reflects how you actually work
What belongs in a CLAUDE.md, what to leave out, why every rule should carry its reason — and why the hard part is keeping it true, not writing it.
Cursor rules: teaching Cursor how you actually work
Where Cursor reads its instructions from, what belongs in a rule, and how to give Cursor the same picture of you that Claude and your other assistants have.
Personal context over MCP: one picture of you for every assistant
Why keeping your working style in one MCP server beats copying it into every tool, and what to check before you let any server hold it: read-only, local, inspectable.