fix(cortex-tui-capture): preserve full Unicode symbols in TUI frame capture

[echobt]

Summary

Fixes an issue where TUI capture frames were displaying incorrectly or empty because only the first character of each cell's symbol was being stored.

Problem

When running with CORTEX_TUI_CAPTURE=1 CORTEX_TUI_CAPTURE_ALL=1 cargo run --bin cortex, the captured frames were not displaying properly because:

1. SnapshotCell stored only a single char instead of the full symbol
2. from_ratatui_buffer used .chars().next() which truncated multi-character grapheme clusters
3. Box-drawing characters, emojis, and other Unicode symbols that span multiple bytes were being cut off

Solution

• Changed SnapshotCell.character: char to SnapshotCell.symbol: String to preserve full symbols
• Updated from_ratatui_buffer to store the complete symbol from each cell
• Updated to_ascii and to_ansi methods to output the full symbol string
• Added helper methods (from_symbol, symbol_with_style, character()) for convenience and backward compatibility

Testing

All 33 tests in cortex-tui-capture pass:

test result: ok. 33 passed; 0 failed; 0 ignored; 0 measured; 0 filtered out

Both cargo check -p cortex-tui-capture and cargo check -p cortex-tui compile successfully.