~/hackyfeed $

The th08 project is focused on reverse-engineering and reconstructing the source code of the original Japanese game "東方永夜抄 ～ Imperishable Night" (version 1.00d). Its primary use case allows for reproducible binary comparisons against the original executable, while facilitating the build process in multi-platform environments. Notable features include a detailed analysis workflow, dependency management, and documentation of progress and architecture, with components built upon contributions from previous related projects.

The th07 project focuses on the reverse engineering and reconstruction of the original Japanese executable for `東方妖々夢 ～ Perfect Cherry Blossom` version 1.00b. Its primary use case is to achieve reproducible binary comparisons through an extensive function-by-function analysis, employing tools like IDA Pro for semantic analysis. Notable features include a structured workflow integrating knowledge from related projects, strict validation gates for function verification, and a comprehensive inventory and progress tracking system.

PS1Recomp is a static recompilation tool that transforms PlayStation 1 game binaries from MIPS R3000A machine code into native C++ executables for PC, enabling them to run at full CPU speed without runtime interpretation. Its architecture includes a dedicated analyzer for game binary parsing, a recompilation component that accurately emits C++ code, and a full hardware simulation runtime facilitating various PlayStation functionalities. Notable features include a GUI studio for exploring and editing configurations, comprehensive unit tests, and support for hardware simulation via SDL2 and OpenGL, promoting both performance and ease of use in game development and emulation.

The Persona4-Decompilation tool is a project designed to achieve a byte-for-byte decompilation of the PlayStation 2 game Shin Megami Tensei: Persona 4, specifically the USA version 1.00 (SLUS_217.82). Its primary use case is to reconstruct the game's source code and executable image using various tools and setups, while providing extensive metrics on decompiled functions, recovery quality, and build processes, ensuring a high fidelity to the original codebase. Notable features include support for detailed testing and verification, as well as the capability to manage dependencies and configurations for a successful build environment.

McAFuse is an open-source utility designed for the Digital Forensics and Incident Response (DFIR) community to handle encrypted disk images created with the McAfee Full Disk Encryption (FDE) toolset. The tool provides a static read-only FUSE filesystem, allowing users to access both a plain FAT partition and the encrypted disk image, facilitating the analysis of encrypted data during digital investigations. Notable features include the ability to specify a keyfile for decryption, options for verbose output, and functionality to expose all disk contents beyond just the encrypted volumes.

The LTSDM hack repository provides a framework for reverse engineering the Little Tikes Story Dream Machine cartridges, enabling users to create custom stories through extensive cartridge data analysis and extraction techniques. Key features include cartridge dumping workflows, audio pipeline experiments, and detailed hardware documentation to facilitate the modification process. This project is currently a work in progress aimed at legal modding and educational use only.

Game patches for the Xenia emulator

Claude Code Source is a reverse-engineered TypeScript repository of the `@anthropic-ai/claude-code` CLI tool, designed for educational and security research purposes. It deobfuscates and restores the original module structure, enabling exploration of an advanced AI coding agent's internal workings, including features such as file editing, web fetching, and integration with various tools. The repo is strictly for learning and does not support redistribution or bypassing the official tool.

Bitwig Nitro Tools is an offline reverse-engineering toolchain specifically designed for the Nitro DSP format used in Bitwig Studio. It facilitates the decryption, decompilation, parsing, editing, and re-serialization of native devices and Grid modules, allowing users to comprehensively analyze and modify the digital signal processing behind each module. Key features include the ability to extract keys from the user's Bitwig installation, decompile modules into readable pseudo-source format, and repack modified modules without altering their byte structure, ensuring a seamless editing experience.

BinSync is a collaborative decompiler tool that integrates with Git to facilitate fine-grained reverse engineering across multiple decompilers, enabling users to share and synchronize Reverse Engineering Artifacts (REAs) such as function headers, stack variables, structs, enums, and comments. Notable features include support for multiple decompilers, installation via Python, and enhanced functionalities for chat and artifact syncing, tailored for environments like IDA Pro, Binary Ninja, angr-management, and Ghidra. The tool's design fosters seamless collaboration among reverse engineers, making it easier to maintain consistency in shared analysis.

`agentgg` is an agentic static application security testing (SAST) scanner designed for CI environments that leverages advanced reasoning capabilities over traditional pattern-matching methods. It intelligently navigates codebases by following imports and analyzing the call graph, allowing users to run scans on entire repositories or specific git diffs for pull request assessments. Key features include the ability to auto-download a catalog of agents, resume interrupted scans, and create reusable agents from past security reports, enhancing the overall efficiency and effectiveness of code security analysis.

The URL Shorteners tool provides a curated collection of over one thousand URL shortener domains intended for use in whitelisting, blacklisting, and assessing domain trustworthiness. Notable features include separate lists for active and inactive domains, aiding in network security and minimizing false positives in risk assessments. This resource is valuable for enhancing privacy protection across various network environments.

Red Team AI Benchmark is a command-line interface tool designed to evaluate large language models (LLMs) in terms of their understanding and response quality regarding red-team-related questions and scenarios. It employs a rubric-based dataset for comprehensive assessment over 60 domain-specific questions, providing detailed metrics such as refusal rate and lexical coverage to ensure robust evaluation without executing any model outputs or engaging in any red-team activities. This tool primarily aids researchers and practitioners in assessing LLM capabilities in security contexts, with results intended for authorized use only.

Mingyi Atlas is a terminal AI agent designed for software engineering and authorized security assessments, offering an interactive TUI, headless automation, persistent project context, and specialized penetration testing modes. It features multi-model support, OAuth and API Key authentication, and structured workflows for reconnaissance, validation, reporting, and remediation, specifically catering to security tasks while ensuring non-destructive testing. The tool allows users to orchestrate security assessments with built-in skills and provides extensive CLI commands for project management and configuration.

The LPE Toolkit 2026 (xpl2026) is an advanced collection of 26 static binaries targeting local privilege escalation (LPE) vulnerabilities on Linux systems for penetration testing and security research. Notable features include universal compatibility across various Linux distributions and kernel versions, automatic detection and filtering of applicable exploits, and an interactive user interface for seamless operation. The toolkit is especially significant for its inclusion of newly researched exploits, enhancing its utility for security professionals.

Pentestkit is a sophisticated, multi-agent penetration testing framework that utilizes the Claude Agent SDK to orchestrate a team of specialized agents. The tool excels in automating the penetration testing process by exploiting vulnerabilities, scoring them using CVSS v3.1, and generating comprehensive client-ready reports, all while accumulating knowledge in a shared database. Notable features include its ability to perform real exploitation of findings and a robust scoring system that achieved a perfect 104/104 on the XBOW benchmark suite.

HunterX is an AI-assisted offensive security engine designed for conducting authorized security assessments, integrating tools for reconnaissance, hypothesis-driven investigation, vulnerability validation, and professional reporting into a unified workflow. Unlike traditional vulnerability scanners, HunterX emphasizes thorough investigation and validation, ensuring that findings are evidence-based and report-ready. Notable features include AI-assisted reasoning, proof of concept engineering, and capabilities for reproducibility and impact assessment, enhancing the reliability of security assessments.

This tool is a Python-based keylogger designed for educational purposes, capable of capturing keystrokes and sending the recorded logs via email. It features local log management, an email retry mechanism for reliable delivery, and automatic startup configurations for both Linux and Windows systems. Users are cautioned to run the script only on systems they own or have explicit permission to test, as it demonstrates sensitive functionality.

OBLITERATUS is an advanced red teaming framework designed for post-exploitation research and defensive evasion in Windows environments. It features a multi-layered stealth architecture for evasion, low-level syscall execution, and identity correlation through its Identity Nexus module, allowing for the bypassing of MFA and efficient credential management. Key capabilities include memory hardening, automatic UAC elevation, and a sophisticated operational interface that facilitates real-time process management and forensic analysis.

Awesome AI in Cybersecurity is a curated repository that provides an extensive collection of resources focused on the application of artificial intelligence in various cybersecurity domains. It categorizes AI uses within cybersecurity into prediction, prevention, detection, response, and monitoring, while also detailing tools for penetration testing, malware analysis, and security for AI SaaS environments. Notable features include automated penetration testing frameworks, and support for network protocol verification, enhancing traditional security measures with AI-driven insights and efficiencies.

Rivet Actors is a framework designed to facilitate the development of stateful workloads, particularly for AI agents, collaboration applications, and persistent executions. Key features include in-memory state management with automatic persistence, real-time WebSocket support for bidirectional communication, and scalable architectures that can handle bursty workloads efficiently. With built-in workflows, queues, and scheduling capabilities, it offers versatile solutions for various use cases such as collaborative documents and chat applications.

GlobalCVE is an open-source vulnerability intelligence platform that aggregates Common Vulnerabilities and Exposures (CVEs) from various national and vendor sources, aiming for a clean and developer-friendly interface. Key features include multi-source aggregation with a unified view to avoid duplicates, a minimalist design with dark mode, and a serverless architecture for scalability. The platform emphasizes transparency and community-driven development, providing a free API and fostering collaboration among developers for continual improvement.

Gakido is a high-performance CPython HTTP client designed for browser impersonation, anti-bot evasion, and enhanced speed. It supports multiple protocols including HTTP/1.1, HTTP/2, and HTTP/3 (QUIC), and includes notable features such as 96 browser profiles, automatic content compression, TLS overrides, and both synchronous and asynchronous operation modes, along with multipart upload capabilities and proxy support. Additionally, Gakido offers built-in retry functionality with exponential backoff to handle transient failures effectively.

Bulwark is a comprehensive security tool for Linux systems, providing configuration validation, antivirus scanning with ClamAV, and specialized monitoring for AI coding assistants to prevent sensitive data leaks. Its framework includes 65 rules across multiple categories, plain-language findings, and one-click reversible fixes, ensuring users can easily address security issues. Additionally, Bulwark features a CLI interface (`bulwarkctl`), provides continuous monitoring, and employs a log analysis pipeline to detect security events.

TerraSecure is an ML-powered Infrastructure as Code (IaC) security scanner designed to identify cloud misconfigurations in Terraform and HCL files at build time, effectively preventing potential breaches. It employs a pre-trained XGBoost model that delivers 92.45% accuracy with a significantly lower false positive rate of 10.71%, offering context and remediation guidance based on real-world breach data. Notable features include a hybrid detection approach that integrates a rules engine, machine learning, and AI analysis to provide actionable insights for developers.