Worked on the DefiLlama-Adapters repository to enhance TVL accuracy for ERC4626 vaults within YieldSeeker. Focused on backend development using JavaScript and blockchain technologies, the work involved implementing on-chain unwrapping of vault shares and robust balance conversion logic. By leveraging on-chain methods such as asset(), totalAssets(), and convertToAssets, the solution addressed vault-specific edge cases and accurately reflected underlying asset values, including complex debt-share scenarios. These improvements increased the integrity and reliability of TVL dashboards, enabling YieldSeeker users to make better-informed risk assessments and business decisions based on more trustworthy DeFi data and analytics.