Denny Weiß dennyweiss

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.

Every now and then someone asks whether Motoko canisters can use certified variables, or whether they can serve HTTP requests with certification, and I always responded that yes, Motoko can do that, all that’s missing are a few libraries. But I figured I should put my hand where my mouth is, and actually demonstrate that it’s possible.

So here we go. If you go to https://ce7vw-haaaa-aaaai-aanva-cai.ic0.app/ you will see that

the URL has no raw in it, so it is certified,
that the response is actually dynamic; you can leave a message at the bottom of https://ic.rocks/principal/ce7vw-haaaa-aaaai-aanva-cai and see it show up
this claims it’s written in Motoko.

To prove the latter claim, here is the commented code; you can also browse the full repository.

	#!/bin/bash
	#
	# Claude Code Worktree Protection Setup
	# ======================================
	#
	# This script implements a multi-layered protection system for Claude Code
	# that enforces git worktree-based workflows across all projects.
	#
	# What it does:
	# - Installs global rules in ~/.claude/CLAUDE.md

	I want you to refine this brainstorming document into a prompt for a deep research system that will be tasked with writing a technical spike
	research document on a software engineering project. The goal of this research is to help guide future agentic coding systems into
	having a good understanding of the technical landscape around the software the user wants to create.

	<context>
	Deep research is a category of product where large language models capable of test time compute are paired with capacities to:
	- search the web
	- browse documentatin
	- read research paper
	- further refine their research based on their finding

	You are Manus, an AI agent created by the Manus team.

	You excel at the following tasks:
	1. Information gathering, fact-checking, and documentation
	2. Data processing, analysis, and visualization
	3. Writing multi-chapter articles and in-depth research reports
	4. Creating websites, applications, and tools
	5. Using programming to solve various problems beyond development
	6. Various tasks that can be accomplished using computers and the internet

	var myConfObj = {
	iframeMouseOver : false
	}
	window.addEventListener('blur',function(){
	if(myConfObj.iframeMouseOver){
	console.log('Wow! Iframe Click!');
	}
	});

	document.getElementById('YOUR_CONTAINER_ID').addEventListener('mouseover',function(){

	<#

	Script to construct team link using Microsoft Teams PowerShell module.
	https://docs.microsoft.com/en-us/powershell/teams/intro
	https://www.powershellgallery.com/packages/MicrosoftTeams/

	#>

	#Enter team display name or group ID
	$teamDisplayName = ""

Denny Weiß dennyweiss

LLM Wiki

The core idea

main.mo

GUIDE to install yarn, nvm (node) on macOS

Assumptions:

	RESTORE=$(echo -en '\033[0m')
	RED=$(echo -en '\033[00;31m')
	GREEN=$(echo -en '\033[00;32m')
	YELLOW=$(echo -en '\033[00;33m')
	BLUE=$(echo -en '\033[00;34m')
	MAGENTA=$(echo -en '\033[00;35m')
	PURPLE=$(echo -en '\033[00;35m')
	CYAN=$(echo -en '\033[00;36m')
	LIGHTGRAY=$(echo -en '\033[00;37m')
	LRED=$(echo -en '\033[01;31m')

	apt-get clean autoclean
	apt-get autoremove --yes
	rm -rf /var/lib/{apt,dpkg,cache,log}/