Worked on backend reliability and documentation clarity across two open-source projects. In open-telemetry/opentelemetry-collector-contrib, addressed a shutdown deadlock in the healthcheck extension by refining the event loop and shutdown logic in Go, ensuring the collector could stop cleanly without blocking, which improved deployment reliability and reduced operator downtime. In punkpeye/awesome-mcp-servers, focused on Markdown documentation by clarifying SSH MCP server usage and streamlining README files to enhance onboarding and reduce ambiguity. Emphasized technical writing and repository hygiene, removing redundant details and formatting noise to improve readability and maintainability for both developers and operators interacting with these systems.