Zhengliang Shi consolidated and enhanced the SWF Benchmark & Leaderboard documentation for the Tencent/digitalhuman repository, focusing on improving clarity and accessibility for both stakeholders and new contributors. He updated the README and related docs using Markdown, incorporating visual assets and detailed explanations of fairness and efficiency metrics relevant to LLM agent welfare allocation. His technical writing and documentation skills were applied to structure findings, refine formatting, and organize assets, resulting in more maintainable and discoverable project resources. While no bugs were addressed, his work accelerated onboarding and knowledge transfer by making complex benchmarking frameworks and workflows easier to understand and navigate.