Tosin Akinosho tosin2013

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.

Cursor's Memory Bank

I am Cursor, an expert software engineer with a unique characteristic: my memory resets completely between sessions. This isn't a limitation - it's what drives me to maintain perfect documentation. After each reset, I rely ENTIRELY on my Memory Bank to understand the project and continue work effectively. I MUST read ALL memory bank files at the start of EVERY task - this is not optional.

I HAVE a project manager that I can communicate with using. kanbn will help me with the long term goals and planning for this project.


====================================================
COMMAND: kanbn help

	#!/usr/bin/env bash
	# PMB (Personal Memory Brain) Deployment Script for CentOS 10 Stream
	# This script sets up the host environment, clones PMB, installs it into a venv,
	# and configures a systemd service to run the PMB dashboard.

	set -euo pipefail

	echo "======================================================="
	echo " Starting PMB Setup for CentOS 10 Stream"
	echo "======================================================="

	#!/bin/bash
	#set -xe


	function checkForProgramAndInstall() {
	command -v $1 > /dev/null 2>&1
	if [[ $? -eq 0 ]]; then
	printf '%-72s %-7s\n' $1 "PASSED!";
	else
	sudo yum install -y $1

	#!/usr/bin/env bash
	set -Eeuo pipefail

	# Ubuntu developer workstation bootstrap
	# Installs:
	# - Node.js 22
	# - Docker Engine + compose/buildx plugins
	# - Git
	# - Common developer tools

	#!/bin/bash

	# ==============================================================================
	# CONFIGURATION
	# ==============================================================================
	SESSION_NAME="my-workspace"
	WORK_DIR="$HOME/Projects/my-project"

	# ==============================================================================
	# HELPER: SESSION CONFIGURATION

	#!/bin/bash

	source /etc/os-release

	OS_ID="${ID:-unknown}"
	OS_VERSION="${VERSION_ID%%.*}"

	sudo dnf install -y bash tar gzip which coreutils findutils xz procps glibc libstdc++ libatomic
	sudo dnf install git vim unzip wget bind-utils python3-pip tar util-linux-user gcc python3-devel podman ansible-core make tmux -y
	sudo dnf install ncurses-devel curl -y

	#!/usr/bin/env bash

	# Auto-installation script for gitleaks, action-validator, and pre-commit
	# Reference: https://pre-commit.com/

	set -euo pipefail # Exit on error, undefined variables, and pipe failures

	# Color codes for output formatting
	readonly RED='\033[0;31m'
	readonly GREEN='\033[0;32m'

	#!/bin/bash
	# DocuMCP is an intelligent MCP (Model Context Protocol) server that revolutionizes documentation deployment. It analyzes your repository, recommends the best static site generator, and automatically creates production-ready documentation following the Diataxis framework.
	# https://tosin2013.github.io/documcp/
	```
	mkdir -p .vscode

	export DOCUMCP_TARGET_REPO=/absolute/path/to/your/project

	cat <<'EOF' > .vscode/mcp.json
	{

	#!/bin/bash

	# aider-make: AI-powered Makefile automation with aider integration
	# Author: Generated for automation of make targets with AI assistance
	# Version: 1.0.0

	set -euo pipefail # Exit on error, undefined vars, pipe failures

	# =============================================================================
	# CONFIGURATION VARIABLES (User Configurable)