Open source.

Laravel Clsx is a utility package designed to streamline the management of HTML class attributes within Laravel applications. It is tailored for developers who want to avoid cluttered template syntax when applying classes based on complex boolean conditions. By porting the popular clsx JavaScript library to PHP, it provides a clean, readable, and consistent API for dynamically concatenating CSS classes in both Blade templates and standard PHP logic.

This repository is a GitHub profile README rather than a runnable codebase — a structured introduction to Bashar Ayyash (Yabasha), an AI systems architect in Amman, Jordan, aimed at employers, collaborators, and anyone evaluating his open-source work. It positions him around a specific thesis: twenty years shipping software, four of them inside a regulated bank building production RAG and agent systems with evals, guardrails, and monitoring in both Arabic and English. The unifying theme is the gap between a working demo and a system that survives real users, and every section of the README serves that framing.

This project provides a rigorous evaluation harness for Arabic speech-to-text, addressing a fundamental problem with standard word error rate (WER) metrics in Arabic: orthographic variation causes naive scoring to penalize correct transcriptions. Arabic has multiple valid spellings for the same spoken word — differences in alef forms, ta marbuta, diacritics, and digit formats — which inflate error rates and can reverse vendor rankings. The tool stratifies results by dialect and code-switching, reports a "normalization gap" as a first-class metric, and separates normalization decisions from recognition accuracy so evaluations remain reproducible and arguable.

@yabasha/cas is a command-line interface tool designed to instantly scaffold professional-grade, AI-ready monorepos. It is tailored for developers building production SaaS applications who require a standardized, high-performance foundation built on the Composable AI Stack. By providing interactive prompts and CLI-driven configuration, the tool allows users to selectively integrate complex architectural components like background workers, external API gateways, and automated evaluation harnesses without manual setup.

The Composable AI Stack (CAS) is a production-ready monorepo template designed for developers building AI-integrated SaaS applications. It provides a structured, modular foundation that abstracts the complexity of scaling from localized prototyping to robust enterprise deployments. By standardizing the integration of LLMs, database interactions, and billing logic, it enables teams to focus on core product features rather than boilerplate architecture.

GEX is a versatile command-line interface tool designed to help developers audit, document, and manage their Node.js package environments. It provides a structured approach to generating reproducible reports for local projects and globally installed packages. By supporting multiple runtime environments—Node.js (npm), Bun, Yarn, and pnpm—it serves as a unified utility for teams needing to track dependency states, enforce license compliance, or simplify vulnerability auditing across diverse infrastructure.

This project provides a comprehensive library of procedural instructions and standard workflows designed to turn AI coding assistants into fully autonomous agile development engines. It is tailored for developers and organizations who want to delegate repository management and project lifecycle tasks—such as sprint planning, backlog grooming, issue triage, and release management—directly to their AI agents. By integrating this skill, users can orchestrate complex tasks like parallel issue resolution and automated retrospectives without leaving their terminal.

This project provides a structured AI copywriting skill designed specifically for B2B SaaS startups. It serves as a comprehensive instructional framework that guides developers and marketers through every stage of the content creation process—from initial intake interviews to final platform-specific optimization. It is intended for anyone looking to move beyond generic AI output by anchoring their marketing copy in established, high-conversion frameworks.

This project provides a robust, scalable architecture for modern full-stack development by combining a Laravel 12 backend with a Next.js 16 frontend in a unified monorepo. It is designed for developers who value explicit structure, clear boundaries between application layers, and high-performance workflows. By organizing the core application code within an apps directory and isolating shared UI primitives in a packages workspace, it enables teams to build complex systems while maintaining a consistent design language.