I'm cosmah, a dev from Uganda

github

Ssekirya Cosmah  • Peerlist

<h3>AI Grant Proposal Software &amp; Research Agent</h3>
Flowspec AI is a powerful AI-powered platform designed to assist NGOs, nonprofits, and teams in creating high-quality funding proposals, contracts, and research studies. It streamlines the entire process, from initial research to final document export.
<ul>
<li>Deep Web Research: Utilizes an AI research agent to conduct in-depth research across the deep web, gathering relevant information for your documents.</li>
<li>Academic Literature Search: Access and search over 300 million academic papers to support your research and findings.</li>
<li>Automated Citations: Generates accurate citations in APA, Harvard, and MLA formats, with DOI verification for academic sources. Includes inline citation support and BibTeX export.</li>
<li>Document Generation: Drafts various professional documents, including grant proposals, contracts, research studies, and more, with tailored compilation logic.</li>
<li>Team Collaboration: Enables team collaboration with role-based access, shared citation libraries, and organization templates for a unified workflow.</li>
<li>Multiple Export Options: Easily export your completed documents to PDF, DOCX (Word), or Markdown formats, ensuring submission-ready files.</li>
<li>Private Knowledge Base: Maintain a secure, project-specific knowledge base for your research and documents.</li>
<li>Grounded AI Writing: Ensures that AI-generated content is based on real sources, reducing the risk of hallucinations and providing evidence-based writing.</li>
</ul>
Flowspec AI is specifically built for professional document creation, offering a distinct advantage over general-purpose AI chatbots by providing isolated project knowledge bases, saved citation libraries, and a structured compile-to-export workflow.

AI Grant Proposal Software & Research Agent

EDR Unhooker is a Windows x64 research tool designed to enhance security research and education. This tool locates ntdll from the PEB, validates PE headers, resolves exports, inspects syscall stubs, and dynamically derives syscall numbers (SSNs). It also applies a minimal PEB anti-debug check. The tool is written in Lwanga and utilizes Linux for building and hosting tests, with Windows x64 as the validation target for the PE binary.<h3>Key Features</h3><ul><li>Locates ntdll from the PEB and validates PE headers.</li><li>Resolves exports and inspects syscall stubs.</li><li>Dynamically derives syscall numbers (SSNs).</li><li>Applies a minimal PEB anti-debug check.</li><li>Cross-builds from Ubuntu/WSL using CMake and lld-link.</li><li>Utilizes gcc-mingw-w64-x86-64 and MinGW import libs for PE linking.</li></ul>This open-source project is available on GitHub, providing a comprehensive platform for developers and researchers to contribute and enhance their security tools.

GitHub - cosmah/edr-unhooker · GitHub

Project Lwanga is a cutting-edge, low-level systems programming language designed with a focus on security research, penetration testing, and shellcode development. Built on the robust LLVM framework, it offers direct system access and minimal runtime overhead, making it ideal for creating security tools and exploits. Key features include encrypted strings, position-independent code, and direct syscalls, ensuring a security-first approach. The language supports inline assembly, naked functions, and register access, providing developers with low-level control. With modern tooling such as an LLVM backend, code formatter, and syntax highlighter, Lwanga is cross-platform, supporting x86_64 and ARM64 on Linux, macOS, and Windows. The project is open source and available under the MIT license, encouraging community contributions and collaboration.<h3>Core Features</h3><ul><li>Security-First: Encrypted strings, position-independent code, direct syscalls.</li><li>Minimal Footprint: Hello World in under 10KB, no libc dependencies.</li><li>Low-Level Control: Inline assembly, naked functions, register access.</li><li>Modern Tooling: LLVM backend, code formatter, syntax highlighter.</li><li>Cross-Platform: x86_64 and ARM64 support on Linux, macOS, and Windows.</li></ul>

Project-Lwanga

Bluetooth thermal printer support for Flutter on Windows. This plugin provides comprehensive features for discovery, pairing, connection, and ESC/POS receipt printing specifically designed for POS applications. It requires Flutter 3.16+ and Dart 3.2+, and is compatible with Windows 10 (1809+) or Windows 11. The plugin supports any Bluetooth thermal printer that uses ESC/POS over SPP (Serial Port Profile).<h3>Key Features</h3><ul><li>Scan for printers and manage connections with ease.</li><li>Print plain text or structured receipts with headers, items, and footers.</li><li>Send raw ESC/POS bytes for advanced printing capabilities.</li><li>Monitor printer status and capabilities through diagnostics.</li><li>Handle errors gracefully with detailed exception handling.</li></ul>This package is ideal for developers looking to integrate thermal printing capabilities into their Windows applications using Flutter. It simplifies the process of connecting and managing Bluetooth printers, ensuring a seamless experience for both developers and end-users.

flutter thermal printer| Flutter package

FlowSpec: Autonomous AI-Powered Testing for React Description: Stop wasting hours writing boilerplate tests. FlowSpec is an AI-powered testing agent that automatically generates, executes, and maintains comprehensive Vitest suites for your React applications. By analyzing your components with GPT-4, it handles the heavy lifting—from edge cases and user interactions to complex mocks—so you can focus on building features.Key Features:<ul><li>Zero Configuration: Auto-installs dependencies and configures TypeScript/Vitest for you.</li><li>Intelligent File Watching: Detects new components or changes in real-time and updates tests instantly.</li><li>Smart Mocks: Automatically generates correct Vitest syntax (vi.fn, vi.mock) and handles explicit TypeScript imports.</li><li>Framework Agnostic: Works seamlessly with Next.js (App &amp; Pages Router) and Vite-based React apps.</li></ul>