~/hackyfeed $

TuneFlow Python SDK (`tuneflow-py`) is a tool designed to enhance music production through AI-driven plugins, allowing users to develop and integrate their own algorithms into various aspects of music creation, such as songwriting and mixing. Its architecture empowers developers to modify any part of the project data model while supporting remote plugin execution for flexible deployment. Notable features include seamless access to all project components and the ability to process complex algorithms, promoting an innovative approach to digital audio workstation capabilities.

Tomato is a steganography tool that embeds hidden messages within natural language text by utilizing minimum-entropy coupling (MEC) code, enabling the creation of undetectable stegotext. Its primary use case is to securely encode and decode plaintext messages while preserving the statistical properties of the cover text, ensuring seamless integration and accurate retrieval of hidden information. Notable features include the use of an LLM for generating coherent cover text and a user-friendly command-line interface for encoding and decoding functionalities.

TimeSide is a scalable audio processing framework written in Python that facilitates extensive audio analysis, transcoding, streaming, and labeling within a high-level API framework. Its notable features include a RESTful API for seamless integration into web applications, asynchronous processing capabilities, and support for various audio formats and advanced feature extraction libraries. Designed for researchers and developers, TimeSide allows for sophisticated prototyping of audio applications and enhances user experiences through metadata enrichment and real-time audio processing.

ThingFinder is a Python tool designed for searching and analyzing both source code and binary executables to identify vulnerabilities, malware, and other patterns. It features a modular interface that allows users to create and integrate custom parsers, enabling tailored analysis based on specific criteria. Additionally, the tool utilizes Ghidra for binary analysis to enhance its ability to understand function call relationships within compiled code.

The "theoretical-physics" repository provides an open-source book on theoretical physics, accessible online and designed for easy building in HTML and PDF formats. Its primary use case is to serve as a comprehensive reference for theoretical physics, featuring tools for building the book using Sphinx and supporting various environments like Conda. Notable features include a structured build process and licensing under the MIT license, enabling users to manage and contribute to the content effectively.

Textbeat is a plaintext music sequencer that allows users to compose music in a column-based format, similar to early music trackers, using a text editor like Vim. Key features include support for various musical techniques such as strumming, arpeggiation, and MIDI CC automation, along with flexible MIDI integration for different operating systems. The tool is designed for musicians seeking a lightweight and scriptable way to create and manipulate music compositions interactively.

TextAttack is a Python framework designed for generating adversarial examples, performing data augmentation, and facilitating model training within Natural Language Processing (NLP) applications. Its primary use case is to enhance the understanding and robustness of NLP models through various adversarial attacks while also enabling dataset augmentation to improve model generalization. Notable features include a command-line interface for executing attacks, support for multiple pre-trained models, and the ability to streamline the model training process with minimal configuration.

The Tegridy MIDI Dataset is a comprehensive collection of high-quality MIDI files aimed at enhancing Music Information Retrieval (MIR) and supporting the development of Music AI models. It includes original and normalized MIDI collections, along with links to extensive external datasets and essential software tools, facilitating various tasks such as audio-to-MIDI transcription and music generation. Notable features of this repository include a wide range of curated datasets, integration of soundfonts for audio rendering, and a robust ecosystem of AI applications, positioned as a unified platform for research in symbolic music processing and generation.

This project consists of an open source library allowing software to connect to data repositories using STIX Patterning, and return results as STIX Observations.

This repository presents a systematic study on code obfuscation techniques aimed at thwarting vulnerabilities detected by LLM-based models. It includes components for managing obfuscated and unobfuscated code, auditing and evaluating LLM vulnerability detection results, and a suite of customizable scripts and tools for code obfuscation and analysis. Notable features include pre-existing datasets across various programming languages, integration with multiple LLM auditor models, and the capability to configure the environment for experimental setups.

SirTunnel is a minimalistic tool that securely exposes local web servers to the public internet via HTTPS using SSH tunneling and the Caddy web server. Its primary use case is to facilitate remote access to development environments without the need for complex server configurations, offering features such as automatic HTTPS certificate management, stateless operation, and compatibility with any standard SSH client. The tool's simplicity allows users to quickly set up secure tunnels while leveraging existing infrastructure.

The singing-transcription-plugin is a Tuneflow Python extension designed for the transcription of singing into musical notation. It utilizes a model derived from the research repository by york135 on singing transcription, enabling users to convert vocal performances into sheet music effectively. Notable features include its integration with Tuneflow and the application of advanced algorithms from recent academic advancements in the field.

SignalBot is a Docker-based messaging bot utilizing Signal CLI, designed for sending messages and facilitating various fun interactions within the Signal messaging platform. It supports multiple commands for generating jokes, weather updates, trivia, and other API-driven content, while operating in both local and non-local executor modes. Key features include dependency on Docker for deployment, integration with external APIs like Giphy, and the ability to run in a containerized or standalone environment.

SecureCode Web is a cybersecurity tool designed for vulnerability detection and secure coding practices across web applications. It leverages datasets incorporating various security frameworks, including OWASP and MITRE, to provide insights on exploitation techniques, risk assessment through CVSS scores, and attack vectors. Notable features include detailed metadata categorization, incident grounding, and integration capabilities for penetration testing and security incident event management (SIEM).

Secure_LLM is a tool designed for fine-tuning large language models, specifically aimed at code generation tasks. Its primary use case involves customizing models such as Salesforce's CodeGen for generating and evaluating code snippets, which includes features for training with LoRA and post-processing generated results for compatibility with CodeQL analysis. Notable features include a structured approach to fine-tuning, code generation, and evaluation using specialized scripts that facilitate integration and performance assessment of the generated code.

SEC-bench is an automated benchmarking framework explicitly designed to evaluate Large Language Model (LLM) agents on real-world software security tasks. Its key features include automated benchmark generation from vulnerability databases, containerized reproducible vulnerability instances, and agent-oriented performance assessments on critical tasks such as vulnerability detection and patching. The tool also offers rich reporting capabilities for enhanced performance tracking and result visualization.

SearXNG is a privacy-focused metasearch engine designed to aggregate search results from various web sources without tracking or profiling users. Its primary use case is to provide users with anonymous search capabilities, while notable features include customizable installation and configuration options for enhanced user control and privacy.

SCAMP is a Python framework for computer-assisted music composition, enabling composers to manage musical time, play back notes through various synthesizers, and export results to music notation formats like MusicXML and Lilypond. It offers features such as flexible playback with custom soundfonts, extensive control over pitch and dynamics, microtonality support, and precise polyphonic tempo management. Designed to balance functionality and artistic freedom, SCAMP facilitates auditioning musical ideas interactively while allowing for personal extensions and modular usage.

Rigging is a versatile LLM framework designed to simplify the integration of language models into production environments. It features structured Pydantic models for input and output management, support for various large language models, and capabilities like prompt definition as Python functions, asynchronous processing, and metadata handling. Notable functionalities include integrated tracing, chat templating, and the ability to manage multiple models through simple connection strings.

ReposVul is a high-quality, repository-level vulnerability dataset designed to support vulnerability detection tasks within software projects. It features a comprehensive framework for data collection, including modules for vulnerability untangling, multi-granularity dependency extraction, and trace-based filtering, allowing for the analysis of 6,134 CVE entries across multiple programming languages. This dataset aids researchers and practitioners in identifying and mitigating vulnerabilities more effectively by providing detailed, organized information on code changes and their interrelations.

REMI (REvamped MIDI-derived events) is a music generation tool that converts MIDI scores into discrete text-like tokens, enabling enhanced modeling of rhythmic patterns with a metrical context. Its key application is the generation of expressive Pop piano compositions using a trained Transformer-XL model, which produces coherent music structures without the need for post-processing. Notable features include controllability over local tempo changes and chord progressions, making it a powerful tool for automated music creation.

ReelPhish is a real-time two-factor phishing tool that enables users to capture credentials by simulating browser interactions through a phishing site. It utilizes Selenium to automate browser actions and can be configured to work with various browsers, including Internet Explorer, Firefox, and Chrome, providing features for customization, such as logging and specific element submission. This tool is primarily used for testing the effectiveness of two-factor authentication measures by simulating phishing attacks.

RedCode is a benchmark tool designed for the evaluation of code agents' capabilities in safely executing and generating code across various programming languages and formats. It features two primary components: RedCode-Exec, which assesses the agents' ability to manage unsafe code execution with over 4,000 test cases, and RedCode-Gen, which evaluates the generation of potentially harmful code through 160 prompts. The tool aims to enhance the safety and reliability of code agents by providing a comprehensive dataset and structured evaluation environment.

ReCode is a benchmark tool designed for the robust evaluation of code generation models, providing over 30 types of semantic perturbations to docstrings, function names, and code. Its primary use case is to assess the resilience and robustness performance of both public and custom code generation models against various natural semantic changes, offering comprehensive analytical comparisons across different models. Key features include automatic dataset perturbation generation, configuration flexibility via `config.json`, and the ability to run evaluations with well-defined metrics, enhancing the usability for model robustness analysis.

The Realtime Audio Feature Server is a low-latency tool that captures and streams live audio features such as band energies and FFT spectra over OSC and WebSocket protocols. Its primary use case is to provide perceptually-tuned audio data to applications like VJ tools and game engines in real time, with notable features including independent onset detection for different frequency bands, user-configurable settings via a browser UI, and optimized processing to ensure minimal latency.