openai/gpt-chat-latest → B (4:1)

Side B makes a substantive behavioral fix by changing rank history to use a 0-based thread_post_index that matches `/t/tag/N` routes, replacing a silent `unwrap_or(0)` fallback with an `expect(...)` to enforce the data invariant, updating rendering, documentation, and integration tests accordingly. Side A is almost entirely Clippy-driven cleanup and minor test maintenance, with its only functional change being adding a missing `#[test]` attribute; most other edits are stylistic or lint-related.