total_lines() was returning sampled_line_count() (only ~300 lines from the initial 64KB scan) during the Loading state, capping the scroll range. Use estimated_lines instead so the user can scroll to any position while indexing runs in the background. get_line() already supports incremental forward scanning on demand. Ultraworked with [Sisyphus](https://github.com/code-yeongyu/oh-my-openagent) Co-authored-by: Sisyphus <clio-agent@sisyphuslabs.ai>