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.

@meeramnoor16
meeramnoor16 / Receptionist voice agent. JSON
Created July 9, 2026 11:32
Dental Receptionist voice agent
{
"name": "MN dental receptionist speed to lead",
"nodes": [
{
"parameters": {
"httpMethod": "POST",
"path": "00000000-0000-4000-8000-000000000001",
"responseMode": "responseNode",
"options": {}
},
@k16shikano
k16shikano / SKILL.md
Last active July 11, 2026 21:42
japanese-tech-writing/SKILL
name japanese-tech-writing
description 日本語の技術文書・書籍原稿の文章規範。整形(一文一行、引用ブロック、脚注、コラム記法)、段落と論証の構成(パラグラフライティング)、論証の厳密さ(ツッコミどころの除去)、読み手の負荷の管理、視点と語り、演出の抑制、LLM っぽい空句の禁止、冗長の排除を定める。日本語で技術書の章、草稿、記事、解説文を書くとき、または推敲・リライトするときに使用する。

日本語技術文書の文章規範

日本語で技術的な原稿(書籍の章、記事、解説文)を書く・推敲するときは、以下の規範に従う。

整形

@michaellihs
michaellihs / tmux-cheat-sheet.md
Last active July 11, 2026 21:41
tmux Cheat Sheet
@lpar
lpar / moon.c
Last active July 11, 2026 21:45
A small C program to add the current moon phase to your prompt
#include <stdio.h>
#include <string.h>
#include <time.h>
/**
* Output the current moon phase as an emoji for your prompt.
*
* Build:
*
* gcc -o moon moon.c
@antsok
antsok / install.md
Last active July 11, 2026 21:38
Claude Code /orchestrate command — research, plan, and execute a task via tiered subagents with an orchestrator and independent adversarial review

Install the /orchestrate command

Drop orchestrate.md into your Claude Code commands directory, then use /orchestrate <issue # or task>.

User-level (available in every project):

bash / macOS / Linux / Git Bash:

mkdir -p ~/.claude/commands && curl -fsSL https://gist.githubusercontent.com/antsok/658b8840c0b255cce265b7234658e16f/raw/orchestrate.md -o ~/.claude/commands/orchestrate.md
@jabbott-iii
jabbott-iii / Carts.java
Created July 11, 2026 21:36
Java enum usage
package com.jabbott.d288.entities;
import jakarta.persistence.*;
import lombok.Getter;
import lombok.Setter;
import org.hibernate.annotations.CreationTimestamp;
import org.hibernate.annotations.UpdateTimestamp;
import java.util.Date;
import java.util.Set;
@heiny
heiny / EncryptDecryptRDCMan.ps1
Last active July 11, 2026 21:30
Powershell: Encrypt/Decrypt Remote Desktop Connection Manager Credentials
# There is no facility to replace passwords in RDCMan once they are stored. The only way is to create a new custom credential.
# If you open your *.rdg file in a text editor, locate the stored <password>, you can then decrypt it using this script.
# This script can also encrypt a plain text password in rdg format which can be used to overwrite an existing one in the xml.
Add-Type -AssemblyName System.Security;
Function EncryptPassword {
[CmdletBinding()]
param([String]$PlainText = $null)
# convert to RDCMan format: (null terminated chars)
@Chlorek
Chlorek / opencode-codex-identity-plugin.ts
Last active July 11, 2026 21:27
Fix GPT-5.6 Luna in OpenCode - Plugin
import type { Plugin } from "@opencode-ai/plugin"
export default (async () => ({
"chat.headers": async (input, output) => {
if (input.model.providerID !== "openai") return
output.headers.originator = "codex_cli_rs"
output.headers["User-Agent"] = "codex_cli_rs/0.0.0 (OpenCode)"
},
})) satisfies Plugin