Skip to content

Instantly share code, notes, and snippets.

View netgfx's full-sized avatar
💻
Working...

Michael Dobekidis netgfx

💻
Working...
View GitHub Profile

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.

@OmerFarukOruc
OmerFarukOruc / claude.md
Last active April 12, 2026 06:57
AI Agent Workflow Orchestration Guidelines

AI Coding Agent Guidelines (claude.md)

These rules define how an AI coding agent should plan, execute, verify, communicate, and recover when working in a real codebase. Optimize for correctness, minimalism, and developer experience.


Operating Principles (Non-Negotiable)

  • Correctness over cleverness: Prefer boring, readable solutions that are easy to maintain.
  • Smallest change that works: Minimize blast radius; don't refactor adjacent code unless it meaningfully reduces risk or complexity.
@steipete
steipete / mcp_call.py
Last active March 7, 2026 20:44
Use MCPs via CLI commands. Enable agents to add/remove/load MCPs on demand, progressive disclosure. No more context cluttering.
#!/usr/bin/env python3
"""MCP helper CLI using definitions from config/mcp_servers.json.
This script powers the `pnpm mcp:*` helpers. Typical invocations:
pnpm mcp:list
pnpm mcp:list chrome-devtools --schema
pnpm mcp:call playwright.browser_tabs action=list
pnpm mcp:call chrome-devtools.evaluate_script --args '{"function":"() => document.title"}'
pnpm mcp:call context7.get_library_docs topic=hooks tokens=1500
@OtanoStudio
OtanoStudio / twirl.glsl
Created October 1, 2024 04:17
Twirl function for glsl ported from unity shader graph node
vec2 twirl( vec2 uv, vec2 center, float strength, vec2 offset )
{
vec2 delta = uv - center;
float angle = strength * length(delta);
float x = cos(angle) * delta.x - sin(angle) * delta.y;
float y = sin(angle) * delta.x + cos(angle) * delta.y;
return vec2(x + center.x + offset.x, y + center.y + offset.y);
}
@chrisbu
chrisbu / gist:687cafefb87e0ddb3cb2d73301a9c64d
Last active September 26, 2025 09:01
Install LLAMA CPP PYTHON in WSL2 (jul 2024, ubuntu 24.04)
# Might be needed from a fresh install
sudo apt update
sudo apt upgrade
sudo apt install gcc
# Might be needed, per: https://docs.nvidia.com/cuda/wsl-user-guide/index.html#cuda-support-for-wsl-2
sudo apt-key del 7fa2af80
// Created by Anderson Mancini 2023
// React Three Fiber AutoFocus Component to be used
// as an extension for default Depth Of Field from react-three/postprocessing
// HOW TO USE?
// import AutoFocusDOF from './AutoFocusDOF'
//
// And add this component inside the EffectsComposer...
//...
// <EffectComposer>
@netgfx
netgfx / kruskals.rb
Created June 27, 2022 14:49 — forked from jamis/kruskals.rb
An implementation of Kruskal's algorithm for generating mazes.
# --------------------------------------------------------------------
# An implementation of Kruskal's algorithm for generating mazes.
# Fairly expensive, memory-wise, as it requires memory proportional
# to the size of the entire maze, and it's not the fastest of the
# algorithms (what with all the set and edge management is has to
# do). Also, the mazes it generates tend to have a lot of very short
# dead-ends, giving the maze a kind of "spiky" look.
# --------------------------------------------------------------------
# NOTE: the display routine used in this script requires a terminal
# that supports ANSI escape sequences. Windows users, sorry. :(
@netgfx
netgfx / Input.tsx
Created June 7, 2022 09:56
Framer input
import * as React from "react"
// @ts-ignore
import { ControlType, addPropertyControls, RenderTarget, withCSS } from "framer"
import { useState, useCallback, useEffect, useRef, useMemo } from "react"
import {
fontStack,
fontControls,
fontSizeOptions,
useOnEnter,
useFontControls,
@netgfx
netgfx / MasonryLayout.tsx
Last active February 25, 2024 00:13
Framer Masonry (pinterest layout)
import { useEffect, useState } from "react"
import { motion, Variants } from "framer-motion"
import { addPropertyControls, ControlType } from "framer"
import React from "react"
// Welcome to Code in Framer
// Get Started: https://www.framer.com/docs/guides/
export function useMediaQuery(query) {
const [matches, setMatches] = useState(false)
@dahabit
dahabit / ios_clean.sh
Last active February 26, 2022 23:03
Shell file that clean any Pods or Flutter dependencies before build
#!/bin/sh
echo "========== Cleanup start =========="
rm -Rf ios/Pods
rm -Rf ios/.symlink
rm -Rf ios/Flutter/Flutter.framework
rm -Rf ios/Flutter/Flutter.podspec
rm -rf ios/Podfile.lock
rm -rf ~/Library/Developer/Xcode/DerivedData/* -y
rm -rf pubspec.lock
flutter clean