Developed and delivered a context-aware skill management feature for the mastra-ai/mastra repository, enabling dynamic skill suggestions based on the request context to improve user experience and agent workflows. Leveraged Node.js and TypeScript to refactor the core skill processor, ensuring available skills are derived directly from requestContext and addressing previously reported issues. Employed test-driven development to validate context-aware skill loading and updated documentation to reflect the new functionality. This work established a stronger foundation for context-aware capabilities, resulting in more relevant skill recommendations and smoother agent interactions, while demonstrating a methodical approach to backend development and comprehensive testing practices.