Skip to content

Instantly share code, notes, and snippets.

View mgeeky's full-sized avatar
💭
Wanna sip a sencha?

Mariusz Banach mgeeky

💭
Wanna sip a sencha?
View GitHub Profile
@HackingLZ
HackingLZ / EDR.md
Created May 4, 2026 22:31
Large EDR Prompt

EDR Reverse Engineering Workspace

Universal workflow for complete reverse engineering of endpoint detection and response (EDR) products. Every EDR dropped into this workspace gets the same systematic treatment: full decompilation, rule extraction, ML model extraction, vulnerability analysis, detection gap analysis, and tradecraft development.

Standardized Folder Structure

When given a new EDR product, create this structure under <product>/:

<product>/
@HackingLZ
HackingLZ / EDR_SKILL.md
Last active September 6, 2026 12:31
EDR Skill
name edr-reverse-engineering
description Standardized workflow for reverse engineering endpoint security products, including extraction, decompilation, rule and model recovery, vulnerability analysis, detection gap analysis, proof-of-concept planning, live probes, and stakeholder reports. Use when Claude is asked to analyze an endpoint detection and response product, endpoint protection agent, security sensor, antivirus engine, or similar endpoint security package from installer artifacts, extracted binaries, live systems, or prior workspace outputs.

EDR Reverse Engineering

Use this skill to run a complete, repeatable reverse-engineering workflow for endpoint security products. Apply the same structure to every product so outputs can be compared across analyses and future work can resume without rediscovery.

@karpathy
karpathy / microgpt.py
Last active September 16, 2026 05:29
microgpt
"""
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

AI Orchestrators Decision Table

An attempt to find a good agent orchestrator or two, that can support these styles of software AI dev:

  1. Mostly AI assisted coding (currently leading: conductor)
  2. Mostly vibe coding (currently exploring configuring pi from scratch)

Ideally one tool could be used for both.

Leave any suggestions in the comments please, all are welcome.

@sherifkozman
sherifkozman / codex-researcher.md
Last active June 22, 2026 15:33
codex-researcher: A Claude Code sub-agent that uses OpenAI Codex CLI for conducting research

name: codex-researcher description: > Use this agent when you need to conduct research on any topic using OpenAI's Codex CLI. This includes gathering information about technologies, companies, market trends, scientific topics, historical events, or any subject requiring in-depth analysis. The agent leverages the codex exec command with web search to perform comprehensive research. model: inherit color: green

@decalage2
decalage2 / olecheck.py
Last active April 23, 2026 09:47
olecheck - a simple script to identify potential CVE-2026-21509 samples
# script to scan MS Office files, looking for "Shell.Explorer" OLE objects which could match CVE-2026-21509
# using oletools - https://github.com/decalage2/oletools
# Philippe Lagadec 2026-01-28
# NOTES:
# According to the MS advisory https://msrc.microsoft.com/update-guide/vulnerability/CVE-2026-21509
# the CVE-2026-21509 vulnerability is related to CLSID "EAB22AC3-30C1-11CF-A7EB-0000C05BAE0B",
# corresponding to the "Shell.Explorer" COM object, which can be used to open the legacy
# Internet Explorer engine (aka Trident/MSHTML) from any application.
# So to exploit CVE2026-21509 from a MS Office document, one could use either an OLE object
@UNC1739
UNC1739 / panw-exclusions.json
Created August 26, 2025 07:14
Default TLS Exclusions for Palo Alto Networks Firewalls
{
"type": "rpc",
"tid": "32",
"action": "PanDirect",
"method": "execute",
"predefinedCacheUpdate": "false",
"cloudCacheUpdate": "false",
"result": {
"result": {
"entry": [
@willccbb
willccbb / grpo_demo.py
Last active September 15, 2026 09:33
GRPO Llama-1B
# train_grpo.py
#
# See https://github.com/willccbb/verifiers for ongoing developments
#
"""
citation:
@misc{brown2025grpodemo,
title={Granular Format Rewards for Eliciting Mathematical Reasoning Capabilities in Small Language Models},
author={Brown, William},

g.co, Google's official URL shortcut (update: or Google Workspace's domain verification, see bottom), is compromised. People are actively having their Google accounts stolen.

Someone just tried the most sophisticated phishing attack I've ever seen. I almost fell for it. My mind is a little blown.

  1. Someone named "Chloe" called me from 650-203-0000 with Caller ID saying "Google". She sounded like a real engineer, the connection was super clear, and she had an American accent. Screenshot.

  2. They said that they were from Google Workspace and someone had recently gained access to my account, which they had blocked. They asked me if I had recently logged in from Frankfurt, Germany and I said no.

  3. I asked if they can confirm this is Google calling by emailing me from a Google email and they said sure and sent me this email and told me to look for a case number in it, which I saw in

#!/usr/bin/env python
# Impacket - Collection of Python classes for working with network protocols.
#
# Copyright Fortra, LLC and its affiliated companies
#
# All rights reserved.
#
# This software is provided under a slightly modified version
# of the Apache Software License. See the accompanying LICENSE file
# for more information.