~/hackyfeed $

Misc-Powershell-Scripts is a collection of various PowerShell scripts designed for diverse tasks and automation. The repository serves as a resource for system administrators and security professionals, featuring tools that facilitate quick execution of common PowerShell functions. Notable features include customizable scripts that can be adapted for specific environments or requirements.

MiroFish is an advanced AI prediction engine that utilizes multi-agent technology to simulate a high-fidelity digital environment, allowing users to predict future scenarios by uploading seed data and specifying prediction requirements in natural language. It features a dynamic interaction of intelligent agents that evolve within this digital sandbox, enabling both serious decision-making and creative exploration without risk. The tool provides detailed prediction reports and serves as a rehearsal platform for policy testing and imaginative scenario development.

`mimikittenz` is a post-exploitation PowerShell tool designed for extracting sensitive data, including plain-text passwords and various types of private information, from the memory of running processes using the Windows `ReadProcessMemory()` function. It supports a wide range of data extraction, including webmail credentials, financial application data, and more, employing customizable regular expressions for targeted information retrieval. Notable features include the ability to target specific processes and the flexibility for users to add custom regex patterns for further data extraction needs.

Mididings is a MIDI router and processor implemented in Python, designed to facilitate the manipulation and routing of MIDI messages in real-time. Its primary use case is audio production and performance, with notable features including support for ALSA and JACK, optional integration for OSC messages, and real-time script modification capabilities. The tool is extensible through various optional libraries and offers a GUI for enhanced user interaction.

mididings is a Python-based MIDI router and processor designed for Linux and macOS, optimized for live performance with capabilities to route, filter, transform, and split MIDI events using a Python domain-specific language (DSL). Notable features include seamless scene switching for dynamic routing configurations, comprehensive MIDI event modification and conversion, as well as monitoring capabilities that allow debugging and interaction with external systems via shell commands or OSC messages. This community-maintained fork ensures compatibility with Python 3 and integrates a modern build system alongside enhanced functionality from previous iterations.

The Midi-Model is a MIDI event transformer designed for symbolic music generation, leveraging advanced tokenization and visualization techniques for enhanced performance. Notable features include an optimized tokenizer with quality filtering for datasets, pretrained model availability, and options for online demonstrations through Hugging Face and Google Colab. The tool is built on PyTorch and includes functionalities for training new models and running music generation applications.

The Microsoft-Extractor-Suite is a comprehensive PowerShell tool designed for collecting and analyzing data from various Microsoft sources, including Unified Audit Logs, Admin Audit Logs, and Entra ID logs. Its notable features include retrieving information about OAuth applications, MFA status, privileged roles, and security alerts, all aimed at enhancing incident response and security analysis in Microsoft environments. The tool is actively maintained and facilitates streamlined extraction of critical operational data to support security assessments and compliance.

MemPalace is a local-first AI memory tool designed for verbatim storage and retrieval of conversation history, utilizing a structured indexing system for efficient semantic searches. It features a pluggable backend architecture, currently defaulting to ChromaDB, allowing users to easily integrate alternative databases without altering the core functionality. The tool prioritizes user privacy by ensuring that data remains on the local machine unless explicitly shared.

Mem0 is an open-source memory layer designed for AI systems that enhances the personalization capabilities of AI assistants and agents by intelligently remembering user preferences and adapting over time. Notable features include single-pass memory extraction, first-class agent-generated facts, entity linking for improved retrieval, multi-signal retrieval methods, and temporal reasoning for context-aware responses. This tool serves primarily in customer support and autonomous systems, optimizing interaction through advanced memory management techniques.

MeetingSummarizer is a desktop utility for recording, transcribing, and summarizing meetings, leveraging OpenAI's Whisper for speech recognition and GPT-3.5-turbo for summarization. The tool features both a command-line interface (CLI) and a graphical user interface (GUI) for ease of use, making it suitable for capturing and synthesizing meeting content on Macbooks with Apple Silicon. Notable features include multilingual transcription and intuitive operation through simple commands or GUI interactions.

The DuckDB / MotherDuck Local MCP Server provides a SQL analytics and data engineering platform for AI assistants and integrated development environments (IDEs), enabling connections to local DuckDB files, in-memory databases, and S3-hosted databases. It supports executing SQL queries, browsing database catalogs, and dynamically switching database connections, allowing for both read and write operations. Notable features include a self-hosted setup that facilitates querying across different databases and a default read-only mode for enhanced security.

mcp-score is an AI-powered music notation tool that facilitates the generation and manipulation of musical scores using natural language input. Its primary use case is to create publication-ready scores for applications like MuseScore, Dorico, and Sibelius, featuring components that include a Python MCP server for live score manipulation, a score generation skill that writes music21 Python scripts, and a MuseScore plugin for real-time interaction with scores. Notable features include the ability to read and arrange existing scores and automate various musical tasks through command execution.

The MCP Gateway & Registry serves as a unified control plane for managing AI assets, including MCP servers, agents, and skills, within an organization. It facilitates centralized discovery, governance, and access control through a single authenticated entry point, replacing the traditional chaotic management of separate assets. Notable features include natural-language search capabilities, a robust audit trail for tracking interactions, and support for deployment on Kubernetes, Amazon ECS, or Docker Compose.

The MCP Gateway & Registry facilitates secure, centralized access to Model Context Protocol (MCP) servers, enhancing AI model connectivity with external systems and tools. It addresses service discovery, governed access, and dynamic tool selection through a unified platform, allowing developers to create smarter agents capable of autonomously discovering and utilizing additional tools. Notable features include a reverse proxy architecture leveraging Nginx, intelligent tool discovery via visual and programmatic interfaces, and support for multiple transport protocols.

ContextForge is an open-source registry and proxy designed to unify MCP, A2A, and REST/gRPC APIs, providing centralized governance, discovery, and observability for AI applications. Key features include a comprehensive tools gateway, support for multi-agent routing, robust API gateway functionalities like rate limiting and authentication, and extensive plugin extensibility with over 40 integrations. It is deployable via PyPI or Docker and supports scalable multi-cluster environments through Redis-backed federation and caching.

MCP CLI is a command-line interface designed for efficient interaction with Model Context Protocol servers, facilitating communication with large language models (LLMs) via integration with tools like CHUK Tool Processor and CHUK-LLM. Notable features include experimental AI virtual memory for context management, advanced execution planning with dry-run capabilities, and the ability to serve interactive HTML applications, ensuring robust session management and production-grade security through automatic secret redaction.

MaybeDont is a library designed to enhance web crawling efficiency by minimizing the downloading of duplicate web pages. Its core feature, the `DupePredictor`, learns to identify significant URL components and predicts duplicate content, allowing crawlers to avoid unnecessary downloads and conserve bandwidth. Additionally, it offers integration with Scrapy through middleware, which employs the predictor to intelligently filter out duplicate requests during the crawling process.

Matchering 2.0 is a containerized web application and Python library designed for audio mastering through reference matching. By inputting a target track and a reference track, it adjusts the target to emulate the reference’s sonic characteristics, including RMS, peak amplitude, and stereo width. Notable features include a new open-source brickwall limiter, improved processing speed and accuracy, and the ease of integration for developers through its extensive library capabilities.

Marimo is a reactive Python notebook designed for reproducibility and ease of collaboration, functioning seamlessly with git and capable of being executed as scripts or deployed as web applications. It enhances productivity by automatically managing cell dependencies, allowing users to interact with UI elements without the need for callbacks, while also offering strong data querying capabilities through SQL. Notable features include its ability to store notebooks as Python files, built-in AI integration, and the option to share work as interactive web apps or presentations.

Maltego TRX is a Python library designed for developing custom transforms to be used within the Maltego environment, facilitating data manipulation and integration for operational intelligence tasks. The library provides a structured project setup, automatic transform discovery, and support for deployment through Docker, while the legacy codebase is no longer actively maintained. Users must utilize Python 3.10 or higher to access the latest features, ensuring compatibility with current software standards.

DARPA Cyber Grand Challenge File Magic

LLM4Decompile is an advanced open-source large language model designed specifically for decompiling binary code into human-readable source code. It utilizes a two-phase process that includes structure recovery and meaningful identifier naming, enhancing the accuracy and usability of the decompiled output. Notably, the tool achieves high re-executability rates and supports various training datasets for improved performance in reverse engineering tasks.

LlamaIndex is an open-source framework designed for constructing and querying data structures that facilitate high-performance interaction with LLMs (large language models). Its primary use case is to create indexes that enhance the retrieval of documents and data in conjunction with LLM-powered applications, allowing for efficient data processing and querying. Notable features include seamless integration with various data sources, support for customizable indexing strategies, and tools for real-time query handling.

The Lidar Calibration Exploratory Tools repository provides a framework for calibrating lidar sensors using a hand-eye method and offers functionalities for building and visualizing calibrated point clouds. Key features include tools for estimating relative transformations between sensors, assessing calibration quality, and visualizing point clouds and sensor poses across Python, C++, and MATLAB implementations. This tool is primarily designed for researchers and engineers working on multi-sensor fusion and environmental mapping applications.

LangGraph is a low-level orchestration framework designed for building, managing, and deploying long-running, stateful agents. It facilitates durable execution with failure recovery, human-in-the-loop capabilities for oversight, and comprehensive memory management, ensuring agents can perform complex workflows reliably. Key features include debugging tools for visibility into agent behavior and production-ready deployment options, making it suitable for sophisticated AI applications integrated within the LangChain ecosystem.