Panic when resuming conversation with Korean characters in todo list

Description

Claude Code panics when trying to resume a conversation (claude -c) that contains Korean (multi-byte UTF-8) characters in the todo list.

Error Message

thread '<unnamed>' (15652) panicked at /rustc/ed61e7d7e242494fb7057f2657300d9e77bb4fcb/library/core/src/str/mod.rs:833:21:
byte index 5 is not a char boundary; it is inside '료' (bytes 3..6) of `완료 `
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace
fatal runtime error: failed to initiate panic, error 5, aborting
zsh: abort      claude -c

Root Cause

The code is attempting to slice a Korean string (완료 meaning "completed") at byte index 5, but this falls in the middle of the character 료 (which occupies bytes 3-6 in UTF-8). Korean characters are 3 bytes each in UTF-8, so byte-based string slicing causes this panic.

Steps to Reproduce

1. Start a Claude Code session
2. Use TodoWrite tool with Korean text (e.g., status or content containing 완료, 진행중, etc.)
3. Exit the session
4. Try to resume with claude -c
5. Application panics

Environment

• OS: macOS (Darwin 24.6.0)
• Claude Code: Latest version via Homebrew
• Shell: zsh

Expected Behavior

Claude Code should properly handle multi-byte UTF-8 characters (Korean, Chinese, Japanese, emoji, etc.) when storing and resuming conversations.

Suggested Fix

Use character-boundary-aware string operations instead of raw byte indexing. In Rust, use .chars() iterator or ensure slicing occurs at valid UTF-8 boundaries.