Skip to content

Instantly share code, notes, and snippets.

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.

@dpaluy
dpaluy / README.md
Last active August 11, 2026 15:49
Download view only protected PDF from Google Drive

Step by step guide to downloading protected PDF from Google Drive

  1. Open the document in Google Docs
  2. Scroll to the bottom of the document, so all the pages are present
  3. Open Developer Tools on separate window and choose the Console tab
  4. Paste the code
  5. Have fun!
@iamaziz
iamaziz / entropy_gain.py
Last active August 11, 2026 15:48
Calculate Entropy and Information Gain for Decision Tree Learning
# -*- coding: utf-8 -*-
# calculating the Entropy and Information Gain for: Learning with Trees
# by: Aziz Alto
# see Information Gain:
# http://www.autonlab.org/tutorials/infogain.html
from __future__ import division
// ==UserScript==
// @name Social Club User Viewer
// @namespace http://tampermonkey.net/
// @version 1.2
// @description Spoofs username check so user pages are visible again
// @author logan
// @match *://socialclub.rockstargames.com/*
// @run-at document-start
// @grant none
// ==/UserScript==
@ossa-ma
ossa-ma / tropes.md
Last active August 11, 2026 15:42
AI Writing Tropes to Avoid — tropes.fyi by ossama.is

AI Writing Tropes to Avoid

Add this file to your AI assistant's system prompt or context to help it avoid common AI writing patterns. Source: tropes.fyi by ossama.is


Word Choice

"Quietly" and Other Magic Adverbs

<?xml version="1.0" encoding="ISO-8859-1"?>
<xsl:stylesheet version="2.0"
xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xmlns:mets="http://www.loc.gov/METS/"
xmlns:dc="http://purl.org/dc/elements/1.1/"
xmlns:dnx="http://www.exlibrisgroup.com/dps/dnx"
xmlns:mods="http://www.loc.gov/mods/v3"
xmlns:dcterms="http://purl.org/dc/terms/"
exclude-result-prefixes="dnx mods xsl mets">
@dinahhandel
dinahhandel / mets.xml
Last active August 11, 2026 15:36
METS to represent the structure of our directory. Includes a tiny bit of PREMIS but mostly media info. For the purposes of validating the structure and contents of our Archival Information Packages. More info on this project here: http://ndsr.nycdigital.org/check-your-aip-before-you-wreck-your-aip/
<?xml version="1.0" encoding="UTF-8"?>
<mets xmlns="http://www.loc.gov/METS/" xmlns:xlink="http://www.w3.org/1999/xlink" xmlns:mi="https://mediaarea.net/mediainfo" xmlns:premis="http://www.loc.gov/premis/v3">
<amdSec ID="amdSec-idm18691448048">
<techMD ID="techMD-idm18691448048">
<mdWrap MDTYPE="PREMIS:OBJECT">
<xmlData>
<premis:object xmlns="http://www.loc.gov/premis/v3" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:type="file" xsi:schemaLocation="info:lc/xmlns/premis-v3 http://www.loc.gov/standards/premis/v2/premis-3-0-draft.xsd" version="3.0">
<objectIdentifier>
<objectIdentifierType>Relative AIP path</objectIdentifierType>
<objectIdentifierValue>objects/PSA_TEST.mov</objectIdentifierValue>
@didley
didley / keymap.json
Last active August 11, 2026 15:35
Zed IDE VIM system copy/paste/save keybindings
[
{
"context": "Editor && !menu",
"bindings": {
"ctrl-c": "editor::Copy",
"ctrl-x": "editor::Cut",
"ctrl-v": "editor::Paste",
},
},
{
@Maciejdziuba
Maciejdziuba / README.md
Last active August 11, 2026 15:34
The Software Factory Playbook — Dex Horthy's 4-gate workflow (Product → Architecture → Program Design → Vertical Slices) as an installable Claude Code skill. From the Dex Horthy episode on David Ondrej's podcast.

The Software Factory Playbook — Dex Horthy's 4-gate workflow as a skill

A Claude Code Agent Skill built from Dex Horthy's (HumanLayer) playbook on David Ondrej's podcast.

"Once the model has written thousands of lines of code, it is harder to change. The sessions that generate design docs are context-light — you get the most model intelligence when you do the hard thinking early."

By default, agents build horizontally: all the backend, then all the frontend, then a 2,000-line diff lands in your lap and reviewing it is your problem. This skill flips that. Every decision that matters gets made before the code exists — where changing your mind costs a sentence, not a rewrite.

What it does