Matt Oldfield asmattic

LLM Wiki

A pattern for building personal knowledge bases using LLMs.

This is an idea file, it is designed to be copy pasted to your own LLM Agent (e.g. OpenAI Codex, Claude Code, OpenCode / Pi, or etc.). Its goal is to communicate the high level idea, but your agent will build out the specifics in collaboration with you.

The core idea

Most people's experience with LLMs and documents looks like RAG: you upload a collection of files, the LLM retrieves relevant chunks at query time, and generates an answer. This works, but the LLM is rediscovering knowledge from scratch on every question. There's no accumulation. Ask a subtle question that requires synthesizing five documents, and the LLM has to find and piece together the relevant fragments every time. Nothing is built up. NotebookLM, ChatGPT file uploads, and most RAG systems work this way.

	{
	"apiVersion": "dashboard.grafana.app/v2beta1",
	"kind": "Dashboard",
	"metadata": {
	"name": "claude-code-metrics",
	"generation": 12,
	"creationTimestamp": "2025-12-10T13:33:56Z",
	"labels": {},
	"annotations": {}
	},

	#!/bin_bash

	# Directory to search for git repositories
	# Defaults to current directory if no argument is provided
	SEARCH_DIR="${1:-.}"

	# Find all .git directories, then go to their parent directory and run git pull
	find "$SEARCH_DIR" -name ".git" -type d -prune \| while IFS= read -r d; do
	REPO_DIR="$(dirname "$d")"
	echo "Pulling changes in: $REPO_DIR"

	<!DOCTYPE html>
	<html lang="en" class="scroll-smooth">
	<head>
	<meta charset="UTF-8">
	<meta name="viewport" content="width=device-width, initial-scale=1.0">
	<title>The Augmented Architect: A Cursor Power User's Guide (2025)</title>
	<script src="https://cdn.tailwindcss.com"></script>
	<script src="https://cdn.jsdelivr.net/npm/chart.js"></script>
	<link rel="preconnect" href="https://fonts.googleapis.com">
	<link rel="preconnect" href="https://fonts.gstatic.com" crossorigin>

	import psutil
	import time
	import datetime
	import csv
	import os

	# --- Configuration ---
	LOG_INTERVAL_SECONDS = 5 # How often to log data (in seconds)
	LOG_DURATION_MINUTES = 60 # How long to run the monitoring (in minutes). Set to 0 for indefinite.
	LOG_FILE_NAME = "system_performance_log.csv"

	#!/bin/bash

	echo -e "\nFind and optionally delete all top level node_modules in a specific directory.\n"

	# Initialize an array to store paths
	NODE_MODULES_PATHS=()

	while IFS= read -r line; do
	size=$(echo "$line" \| awk '{print $1}')
	path=$(echo "$line" \| awk '{print $2}')

	# Add the 3rd party PPA
	sudo add-apt-repository ppa:krifa75/eog-ordissimo

	# Update package cache and install library
	sudo apt update && sudo apt install webp-pixbuf-loader

	# remove the Ubuntu PPA, because it also contains the variant
	# of eog (Eye of GNOME, GNOME’s default image viewer) that
	# might break your other packages
	sudo add-apt-repository --remove ppa:krifa75/eog-ordissimo

	/** Entry header */
	.site-content .entry-header .entry-title {
	text-align: left;
	}

	.entry-title {
	margin: 0 0 0 10px;
	}

	/** fil up more of page */


	sudo apt install snapd && sudo snap install snap-store && sudo reboot

	{
	"files.watcherExclude": {
	"/.git/objects/": true,
	"/.git/subtree-cache/": true,
	"/coverage/": true,
	"*/node_modules//**": true,
	"/server/data/": true,
	"/dist/": true,
	"/modules/": true,
	"/info_gathering/": true,