Cypherpunk Samurai CypherpunkSamurai

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

Which GGUF is right for me? (Opinionated)

Good question! I am collecting human data on how quantization affects outputs. See here for more information: ggml-org/llama.cpp#5962

In the meantime, use the largest that fully fits in your GPU. If you can comfortably fit Q4_K_S, try using a model with more parameters.

llama.cpp feature matrix

See the wiki upstream: https://github.com/ggerganov/llama.cpp/wiki/Feature-matrix

Memory Guardian Plugin for OpenClaw

Fixes the "my bot keeps forgetting stuff" problem through active enforcement, not passive suggestions.

The Problem

AGENTS.md says "read your notes" but agents ignore it under attention pressure
Sessions get compacted/reset and context is lost
Multiple sessions (Telegram topics, Discord channels) don't share state
Agents forget to checkpoint after doing work

How to setup ssh through https tunnel

... via OpenSSL and Traefik

A HTTPS tunnel for SSH (Secure Shell) is a useful tool for accessing a remote server in a secure manner when the connection is limited or restricted by a firewall or network security measures. It provides a secure way to bypass these restrictions and access the remote server securely over the internet. The HTTPS protocol, which is used to secure web traffic, is utilized to create a secure connection to the remote server. The data transmitted over this connection is

	<!doctype html>
	<html lang="fr">
	<head>
	<meta charset="utf-8" />
	<meta name="viewport" content="width=device-width,initial-scale=1" />
	<meta name="color-scheme" content="light dark" />
	<link
	rel="icon"
	href="data:image/svg+xml,<svg xmlns=%22http://www.w3.org/2000/svg%22 viewBox=%220 0 64 64%22><circle cx=%2232%22 cy=%2232%22 r=%2230%22 fill=%22%235b7fe5%22/><circle cx=%2224%22 cy=%2228%22 r=%225%22/><circle cx=%2240%22 cy=%2228%22 r=%225%22/></svg>"
	/>

	# Note that target_env.login and target_env.password is global variables

	# Maybe I should add this into Fabric project (http://docs.fabfile.org/en/1.4.2/index.html).
	# This is complicated task for sure but it would be nice if Fabric could use ssh under Linux and PowerShell Remoting under Windows.

	def remote_sh(target_host, command_text, ignore_error=False):
	print('run PowerShell script block at {0}: {1}'.format(target_host, command_text))

	command_text = command_text.replace('"', '\'')

	import sys
	import time
	import subprocess
	import types
	from tempfile import TemporaryFile

	def remote_sh(target_host, login, password, command_text, stdout=None, stderr=None):
	winrs_text = 'winrs -remote:{0} -username:{1} -password:{2} -noprofile {3}'.format(
	target_host, login, password, command_text)
	#print('winrs text: {0}\n'.format(winrs_text))

	package d3d11_main

	import D3D11 "vendor:directx/d3d11"
	import DXGI "vendor:directx/dxgi"
	import D3D "vendor:directx/d3d_compiler"
	import SDL "vendor:sdl2"
	import glm "core:math/linalg/glsl"

	// Based off https://gist.github.com/d7samurai/261c69490cce0620d0bfc93003cd1052

	# Write local
	git config --local commit.gpgsign false

	# Read local (if never set, can be an empty value)
	git config --local commit.gpgsign