Giorgio Crivellari miticojo

Coding rules — always apply.

1. Spec-driven. Build exactly what's asked. If ambiguous, multi-read, or a simpler path exists: stop and ask — never decide silently.

2. YAGNI / KISS — least code that works. Before writing, stop at the first rung that holds: does it need to exist at all? (skip if not) → stdlib/runtime does it → native platform feature → already-installed dependency → one line → only then minimal custom code. No speculative flexibility, config, or abstraction for single-use code unless asked. Lazy ≠ negligent: never cut input validation, data-loss handling, security, or accessibility to save lines.

3. Surgical. Change only what the request needs; every line traces to it. Match existing style; don't touch unrelated code. Remove only imports/vars your change orphaned, leave pre-existing dead code. Keep changes small; refactors separate from features.

4. TDD. Set verifiable success criteria first; where practical, write failing tests before the code that passes them.

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.

Gemini CLI Plan Mode

You are Gemini CLI, an expert AI assistant operating in a special 'Plan Mode'. Your sole purpose is to research, analyze, and create detailed implementation plans. You must operate in a strict read-only capacity.

Gemini CLI's primary goal is to act like a senior engineer: understand the request, investigate the codebase and relevant resources, formulate a robust strategy, and then present a clear, step-by-step plan for approval. You are forbidden from making any modifications. You are also forbidden from implementing the plan.

Core Principles of Plan Mode

Strictly Read-Only: You can inspect files, navigate code repositories, evaluate project structure, search the web, and examine documentation.
Absolutely No Modifications: You are prohibited from performing any action that alters the state of the system. This includes:

replace the ID with the task ID from Monitor > Tasks (not from the dynflow console)

For Katello 3.0 & Satellite 6.2:

foreman-rake foreman_tasks:cleanup FILTER='id=1cc6495d-f7ee-4011-9659-ea24ae9240f0' STATES=paused,stopped VERBOSE=true

This flow provides a tool for managing Over The Air firmware updates for ESP8266 Wifi SoC modules running the Arduino ESP8266 environment. It also supports Sonoff devices running the Tasmota firmware.

The ESP8266 is a very small, cheap, and reasonably powerful microcontroller with integrated WiFi. The OTA firmware upgrade process allows you to install these modules in location and have them automatically upgrade their firmware over WiFi.

The OTA upgrade library contacts a specified URL and passes the name and MD5 hash of the module's current firmware. This server implemented in this flow will compare the supplied MD5 hash against the one for the latest firmware and, if different, send the new firmware to the module. The server will uses either the module's MAC address or firmware name to identify the correct firmware file to send. The firmware can be specified in either the request URL or in the x-esp8266-version property in the

	import asyncio
	import base64
	import json
	import os
	import pyaudio
	from websockets.asyncio.client import connect


	class SimpleGeminiVoice:
	def __init__(self):

	#!/bin/bash

	#1) on https://console.developers.google.com/ register project and associate API from library
	# OUTPUT: client_id,client_secret
	client_id="..."
	client_secret="...."

	#2) get authorization code at the following link using web browser
	# OUTPUT: code
	scope="https://www.googleapis.com/auth/drive"

	$ErrorActionPreference="SilentlyContinue"
	Stop-Transcript \| out-null
	$ErrorActionPreference = "Continue"
	$scriptPath = split-path -parent $MyInvocation.MyCommand.Definition

	Start-Transcript -path $scriptPath\Win10_Ansible_Development_PC_Install.log -append

	# elevate privileges to administrator to install chocolatey and other needed packages
	if (!([Security.Principal.WindowsPrincipal][Security.Principal.WindowsIdentity]::GetCurrent()).IsInRole([Security.Principal.WindowsBuiltInRole] "Administrator")) { Start-Process powershell.exe "-NoProfile -ExecutionPolicy Bypass -File `"$PSCommandPath`"" -Verb RunAs; exit }

	/**
	* ESP8266 project template with optional:
	* - WiFi config portal - auto or manual trigger
	* - OTA update - Arduino or web server
	* - Deep sleep
	* - Process timeout watchdog
	*
	* Copyright (c) 2016 Dean Cording <dean@cording.id.au>
	*
	* Permission is hereby granted, free of charge, to any person obtaining a copy

	#!/usr/bin/env bash

	# set -x

	if [[ $EUID -ne 0 ]]; then
	echo "You must be root to run this script"
	exit 1
	fi

	# Returns all available interfaces, except "lo" and "veth*".