fix: UTF-8 Panic bei String-Truncation (db.rs + knowledge.rs) [appimage]

Crash: "byte index 200 is not a char boundary; inside '─' (bytes 198..201)"
Fix: char_indices() statt byte-slice für sicheres Abschneiden bei
Multi-Byte-Zeichen (ä, ü, ─, etc.)

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

src-tauri/src/db.rs

@@ -717,7 +717,13 @@ impl Database {
         let mut summary_parts: Vec<String> = Vec::new();
         for (_, role, content) in &old_messages {
             let preview = if content.len() > 200 {
-                format!("{}...", &content[..200])
+                // Sicher an Char-Boundary abschneiden (Multi-Byte wie ─, ä, ü)
+                let end = content.char_indices()
+                    .take_while(|(i, _)| *i < 200)
+                    .last()
+                    .map(|(i, c)| i + c.len_utf8())
+                    .unwrap_or(200.min(content.len()));
+                format!("{}...", &content[..end])
             } else {
                 content.clone()
             };

src-tauri/src/knowledge.rs

@@ -993,9 +993,14 @@ pub async fn format_tool_hints(
 
     for entry in entries {
         hints.push(format!("\n**{}** ({})", entry.title, entry.category));
-        // Content auf ~300 Zeichen kürzen
+        // Content auf ~300 Zeichen kürzen (sicher an Char-Boundary)
         let content = if entry.content.len() > 300 {
-            format!("{}...", &entry.content[..300])
+            let end = entry.content.char_indices()
+                .take_while(|(i, _)| *i < 300)
+                .last()
+                .map(|(i, c)| i + c.len_utf8())
+                .unwrap_or(300.min(entry.content.len()));
+            format!("{}...", &entry.content[..end])
         } else {
             entry.content
         };