Skip to content

Instantly share code, notes, and snippets.

View ruvnet's full-sized avatar
💭
hacking the multiverse.

rUv ruvnet

💭
hacking the multiverse.
View GitHub Profile
@ruvnet
ruvnet / polymorphic-obfuscation.md
Last active November 16, 2024 16:01
This example demonstrates basic polymorphic obfuscation techniques, including encryption, variable code structure, and behavioral adaptation.

Polymorphic obfuscation algorithm involves several complex techniques to ensure the code changes its form each time it is executed, while maintaining its original functionality. Here’s a detailed overview and an example of how you might implement such an algorithm.

Key Techniques in Polymorphic Obfuscation

  1. Code Obfuscation:

    • Use encryption, compression, or other obfuscation methods to conceal the code's true nature.
    • Example: Encrypt the main body of the code and add a decryption function that decrypts the code before execution[3].
  2. Dynamic Encryption Keys:

  • Use different encryption keys for each new instance of the code.
@ruvnet
ruvnet / power-automate.md
Created November 16, 2024 15:07
This project integrates OpenAI's GPT-4o large language model with Power Automate Desktop to create an advanced AI-powered automation system. It uses real-time streaming via WebSockets to enable the AI to observe and interact with your desktop, allowing for dynamic and intelligent automation of tasks.

Introduction:

This project integrates OpenAI's GPT-4o large language model with Power Automate Desktop to create an advanced AI-powered automation system. It uses real-time streaming via WebSockets to enable the AI to observe and interact with your desktop, allowing for dynamic and intelligent automation of tasks.

Key Features:

  • Real-time desktop streaming to GPT-4o via WebSockets
  • AI-powered analysis and decision making for desktop automation
  • Easy setup with guided installation script
  • Customizable automation actions and workflows
  • Seamless integration with Power Automate Desktop
@ruvnet
ruvnet / captcha-bot.md
Created November 16, 2024 14:32
automate the process of solving CAPTCHAs using a desktop computer, you can combine several approaches involving automation tools

To automate the process of solving CAPTCHAs using a desktop computer, you can combine several approaches involving automation tools, CAPTCHA solving services, and scripting. Here’s a detailed guide on how to set this up:

Using Automation Tools

Power Automate Desktop

You can use Power Automate Desktop (formerly Microsoft Power Automate Desktop) to automate the interaction with the CAPTCHA. Here’s a general outline based on the videos and descriptions provided:

  1. Capture CAPTCHA Image:
    • Use Power Automate Desktop to capture a screenshot of the CAPTCHA challenge.
  • Extract the CAPTCHA image from the screenshot[1][4].
@ruvnet
ruvnet / implementation.md
Created November 16, 2024 14:25
Test-Time Training is an innovative approach in machine learning that allows models to adapt and improve their performance during the testing phase

Basic PyTorch implementation of Llama 3.2:

import torch
import torch.nn as nn
import torch.nn.functional as F
import math

class RMSNorm(nn.Module):
    def __init__(self, dim, eps=1e-6):

Auto-Fixer Script

Introduction

The Auto-Fixer script is a powerful tool designed to automatically test and fix React components in a project. It leverages the London school of Test-Driven Development (TDD) and uses an AI-powered code assistant to iteratively improve failing tests and component code.

Features

  • Automatically runs tests for specified React components
  • Analyzes test failures and error messages
@ruvnet
ruvnet / AIDO.md
Last active November 16, 2024 08:43
AI-Driven Decentralized Organization (AIDO) using Supabase and LangChain.js. It includes all necessary files, configurations, tests, and deployment components to get started. Feel free to extend and customize the system based on your specific requirements.

AI-Driven Decentralized Organization (AIDO)

In the evolving landscape of organizational structures, the AI-Driven Decentralized Organization (AIDO) emerges as a groundbreaking concept that leverages artificial intelligence to enable autonomous and distributed decision-making without relying on traditional blockchain or cryptocurrency frameworks. Unlike conventional Decentralized Autonomous Organizations (DAOs) that depend on blockchain technology for transparency and security, AIDO harnesses a network of specialized AI agents to collaboratively manage operations, make strategic decisions, and allocate tasks efficiently.

At the core of AIDO lies a diverse set of AI agents, each equipped with unique expertise in areas such as finance, operations, strategy, and human resources. These agents interact through a sophisticated distributed decision-making system, where proposals are generated, evaluated, and voted upon using advanced machine learning algorithms and natural language processing. A robust consensus

Crypto Trading Bot with Narrative Prediction Strategy

Introduction

This project implements a profit-taking crypto trading bot specifically designed for Crypto.com. By leveraging narrative-based predictions using liteLLM, this bot aims to enhance exit and reinvestment decisions. Recent research has shown that narrative forecasting—using storytelling techniques to frame predictions—significantly improves the accuracy of language model predictions in financial forecasting, with up to an 80% accuracy improvement in some scenarios. This approach combines both technical analysis and narrative-based market sentiment for a unique and optimized trading strategy.

Approach

  1. Purchase Price Tracking: This script maintains records of purchase prices for each asset in a structured database using SQLite and SQLAlchemy.
  2. API Integration: The bot interacts with Crypto.com’s API to retrieve real-time market data and execute buy/sell orders.
@ruvnet
ruvnet / Prediciton.md
Last active November 14, 2024 07:09
A prediction framework & prompt uses a "future retrospective" approach where predictions are framed as historical analysis from a future date. This method has proven particularly effective for economic indicators, market trends, and event outcomes when combined with rigorous backtesting and statistical validation

Predictive Narrative Framework & Prompt

This framework leverages research from Baylor University showing that language models achieve significantly higher accuracy when making predictions through narrative storytelling rather than direct forecasting.

Research

How to structure prompts using the narrative approach that proved more successful than direct prediction:

Direct Prompt (Less Effective)

@ruvnet
ruvnet / subliminal.md
Last active November 12, 2024 01:39
A detailed algorithm and implementation framework for the 25th frame subliminal message technique:

A detailed algorithm and implementation framework for the 25th frame subliminal message technique:

The 25th Frame Subliminal Algorithm is a technique for embedding subliminal messages into video content based on human visual perception limitations. Here's a brief overview:

Core Concept

The algorithm works by inserting an additional frame containing a subliminal message into a standard video sequence that typically runs at 24 frames per second[1]. This extra frame is displayed for less than 50 milliseconds, making it imperceptible to conscious awareness while still being processed by the subconscious mind[3].

Key Benefits

  • Subconscious Processing: The inserted frames bypass conscious perception while still being registered by the brain[1][2].
  • Measurable Impact: Studies have shown physiological responses through EEG measurements, pulse monitoring, and blood oxygen levels when exposed to these frames[1].
@ruvnet
ruvnet / agentic.md
Last active November 10, 2024 15:18
agentic.js: A modern, secure, and scalable JavaScript/TypeScript library designed for Deno and Node.js environments. It provides a unified interface to interact with multiple AI language model providers like OpenAI and Anthropic. Incorporating agentic capabilities through LangChain.js integration, agentic.js allows for advanced AI applications, …

💍 agentic.js

One Ai library to rule them all

agentic.js is a modern, secure, and scalable JavaScript/TypeScript library designed for Deno and Node.js environments.

It provides a unified interface to interact with multiple AI language model providers like OpenAI and Anthropic. Incorporating agentic capabilities through LangChain.js integration, agentic.js allows for advanced AI applications, including custom tool creation and dynamic agent architectures.

The library supports multiple models and providers using dynamic configurations and offers optional support for PostgreSQL/Supabase as a database backend for caching and data storage.

Key Features