Duncan Ogilvie mrexodia

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.

Terminals should generate the 256-color palette from the user's base16 theme.

If you've spent much time in the terminal, you've probably set a custom base16 theme. They work well. You define a handful of colors in one place and all your programs use them.

The drawback is that 16 colors is limiting. Complex and color-heavy programs struggle with such a small palette.

name

plant-seed

Plant Seed Command

Plant ideas you want to tend - instant capture from context, with optional enrichment.

Agent Guidelines for Rust Code Quality

This document provides guidelines for maintaining high-quality Rust code. These rules MUST be followed by all AI coding agents and contributors.

Your Core Principles

All code you write MUST be fully optimized.

"Fully optimized" includes:

Agent Guidelines for Python Code Quality

This document provides guidelines for maintaining high-quality Python code. These rules MUST be followed by all AI coding agents and contributors.

Your Core Principles

All code you write MUST be fully optimized.

"Fully optimized" includes:

description

Interview me about the plan

argument-hint

plan

model

opus

Read this plan file $1 and interview me in detail using the AskUserQuestionTool about literally anything: technical implementation, UI & UX, concerns, tradeoffs, etc. but make sure the questions are not obvious.

	import os,math,random,argparse
	parser=argparse.ArgumentParser()
	parser.add_argument('--n_embd',type=int,default=16)
	parser.add_argument('--n_layer',type=int,default=1)
	parser.add_argument('--block_size',type=int,default=8)
	parser.add_argument('--num_steps',type=int,default=1000)
	parser.add_argument('--n_head',type=int,default=4)
	parser.add_argument('--learning_rate',type=float,default=1e-2)
	parser.add_argument('--seed',type=int,default=42)
	args=parser.parse_args()

	"""
	The most atomic way to train and run inference for a GPT in pure, dependency-free Python.
	This file is the complete algorithm.
	Everything else is just efficiency.

	@karpathy
	"""

	import os # os.path.exists
	import math # math.log, math.exp

	#!/usr/bin/env python3
	"""
	Codebase analyzer using Recursive Language Models (RLM) via DSPy.

	Based on: https://kmad.ai/Recursive-Language-Models-Security-Audit

	Usage:
	python analyze-codebase.py --mode security --output report.md
	python analyze-codebase.py --mode documentation --exclude tests,vendor
	python analyze-codebase.py --mode quality --max-iterations 50

	---
	argument-hint: [path to source image]
	description: Generate favicons from a source image
	---

	Generate a complete set of favicons from the source image at `$1` and update the project's HTML with the appropriate link tags.

	## Prerequisites

	First, verify ImageMagick v7+ is installed by running: