fix: consolidated UTF-8 safety improvements for string slicing

[echobt]

Summary

This PR consolidates 5 UTF-8 safety fixes into a single, cohesive change.

Included PRs:

• fix(cli): use safe UTF-8 slicing in import command base64 extraction #31: Use safe UTF-8 slicing in import command base64 extraction
• fix(notifications): use safe UTF-8 slicing for session IDs #32: Use safe UTF-8 slicing for session IDs in notifications
• fix(resume): use char-aware string truncation for UTF-8 safety #33: Use char-aware string truncation for UTF-8 safety in resume
• fix(cli): use safe UTF-8 slicing for session IDs in lock command #35: Use safe UTF-8 slicing for session IDs in lock command
• fix(agents): validate UTF-8 boundaries in mention parsing #37: Validate UTF-8 boundaries in mention parsing

Changes:

• Replaced direct byte slicing with char-aware methods
• Added floor_char_boundary checks before slicing
• Prevents panics from slicing multi-byte characters

Files Modified:

• src/cortex-agents/src/mention.rs
• src/cortex-cli/src/import_cmd.rs
• src/cortex-cli/src/lock_cmd.rs
• src/cortex-cli/src/utils/notification.rs
• src/cortex-resume/src/resume_picker.rs

Closes #31, #32, #33, #35, #37

[greptile-apps]

Greptile Overview

Greptile Summary

This PR consolidates 5 separate UTF-8 safety fixes into a unified improvement, preventing potential runtime panics when slicing strings containing multi-byte characters (emojis, CJK characters, etc.).

Key Changes:

• mention.rs: Added safe_slice_up_to() and safe_slice_from() helper functions that check is_char_boundary() and search for valid boundaries before slicing
• import_cmd.rs: Replaced direct slicing with .get() method for safe Option-based access when extracting base64 data from user messages
• lock_cmd.rs: Introduced safe_char_prefix() to truncate session IDs by character count (not bytes) using char_indices()
• notification.rs: Rewrote session ID truncation logic to use char_indices() for safe multi-byte handling
• resume_picker.rs: Updated truncate_string() to count characters using .chars() iterator instead of byte length

Testing:
All changes include comprehensive unit tests covering ASCII, multi-byte UTF-8 (Japanese characters - 3 bytes each), emojis (4 bytes each), and mixed content.

Impact:
Eliminates panic risks in edge cases where session IDs, user inputs, or file paths contain multi-byte UTF-8 characters.

Confidence Score: 5/5

• This PR is safe to merge with no identified risks
• All changes follow established Rust UTF-8 safety patterns, include comprehensive test coverage for edge cases, and address real panic risks. The implementations use standard library methods (is_char_boundary, char_indices, .get()) correctly.
• No files require special attention

Important Files Changed

Filename	Overview
src/cortex-agents/src/mention.rs	Added helper functions safe_slice_up_to and safe_slice_from to handle UTF-8 boundaries safely, with comprehensive tests for multi-byte characters
src/cortex-cli/src/import_cmd.rs	Replaced direct string slicing with .get() method for safe UTF-8 boundary handling in base64 extraction logic
src/cortex-cli/src/lock_cmd.rs	Introduced safe_char_prefix function to safely truncate session IDs by character count instead of byte count, with comprehensive tests
src/cortex-cli/src/utils/notification.rs	Rewrote session ID truncation using char_indices() to safely handle multi-byte UTF-8 characters in notifications
src/cortex-resume/src/resume_picker.rs	Updated truncate_string to use character-based truncation with .chars() iterator, plus comprehensive UTF-8 tests

Sequence Diagram

sequenceDiagram
    participant Client as Client Code
    participant OldAPI as Old (Unsafe) Slicing
    participant NewAPI as New (Safe) UTF-8 Slicing
    participant String as Rust String

    Note over Client,String: Before: Direct byte-based slicing
    Client->>OldAPI: &text[..pos]
    OldAPI->>String: Slice at byte position
    alt Byte pos is mid-character
        String-->>OldAPI: ❌ PANIC!
        OldAPI-->>Client: Thread crash
    else Byte pos is valid boundary
        String-->>OldAPI: ✅ Valid slice
        OldAPI-->>Client: Return slice
    end

    Note over Client,String: After: Character-boundary aware slicing
    Client->>NewAPI: safe_slice_up_to(text, pos)
    NewAPI->>String: Check is_char_boundary(pos)
    alt Position invalid
        NewAPI->>NewAPI: Search backwards for valid boundary
        NewAPI->>String: Slice at valid boundary
        String-->>NewAPI: ✅ Valid slice
    else Position valid
        NewAPI->>String: Slice at position
        String-->>NewAPI: ✅ Valid slice
    end
    NewAPI-->>Client: Always returns valid UTF-8

    Note over Client,String: Alternative: Using .get() method
    Client->>NewAPI: text.get(start..end)
    NewAPI->>String: Option-based safe access
    alt Invalid boundaries
        String-->>NewAPI: None
        NewAPI-->>Client: Handle gracefully
    else Valid boundaries
        String-->>NewAPI: Some(&str)
        NewAPI-->>Client: Use slice
    end

Loading

[greptile-apps]

_{5 files reviewed, no comments}

_{Edit Code Review Agent Settings | Greptile}

[echobt]

Closing this PR to consolidate into a single mega-PR combining all bug fixes. The changes will be included in a new consolidated PR.