Improve Chinese character support with robust Unicode property union

Replace hard-coded Unicode range with comprehensive Unicode property approach
to fix incomplete Han character coverage in MCP tool name formatting.

Changes:
- Replace \u4e00-\u9fa5 range with union of Unicode Script and Block properties
- Use \p{IsHan} + \p{InCJK_Unified_Ideographs} + \p{InCJK_Compatibility_Ideographs}
- Fix boundary case where \u9fff was incorrectly excluded by script-only approach
- Add comprehensive test coverage for all Han character blocks and edge cases

Technical details:
- Addresses Unicode Script vs Block classification differences across JDK versions
- \u9fff (鿿) is in CJK Unified Ideographs block but not Han script in some JDKs
- Union approach ensures complete coverage while maintaining exclusion of other scripts
- Future-proof solution that automatically includes new Han characters in Unicode updates

Test coverage added:
- CJK Unified Ideographs boundary cases (\u4e00, \u9fff)
- CJK Extension A characters (\u3400)
- CJK Compatibility Ideographs (\uf900)
- Mixed character block scenarios
- Proper exclusion verification for non-Han scripts (Hiragana, Emoji, etc.)

Fixes incomplete Chinese character support while maintaining backward compatibility
and minimal risk profile of the original change.

Signed-off-by: shishuiwuhen2009
Signed-off-by: Mark Pollack <[email protected]>

Auto-cherry-pick to 1.0.x
Fixes spring-projects#4192

Author: markpollack

mcp/common/src/main/java/org/springframework/ai/mcp/McpToolUtils.java

@@ -80,8 +80,10 @@ public static String prefixedToolName(String prefix, String toolName) {
 		String input = prefix + "_" + toolName;
 
 		// Replace any character that isn't alphanumeric, underscore, or hyphen with
-		// concatenation
-		String formatted = input.replaceAll("[^a-zA-Z0-9_-]", "");
+		// concatenation. Support Han script + CJK blocks for complete Chinese character
+		// coverage
+		String formatted = input
+			.replaceAll("[^\\p{IsHan}\\p{InCJK_Unified_Ideographs}\\p{InCJK_Compatibility_Ideographs}a-zA-Z0-9_-]", "");
 
 		formatted = formatted.replaceAll("-", "_");
 

mcp/common/src/test/java/org/springframework/ai/mcp/ToolUtilsTests.java

@@ -93,6 +93,121 @@ void prefixedToolNameShouldThrowExceptionForNullOrEmptyInputs() {
 			.hasMessageContaining("Prefix or toolName cannot be null or empty");
 	}
 
+	@Test
+	void prefixedToolNameShouldSupportChineseCharacters() {
+		String result = McpToolUtils.prefixedToolName("前缀", "工具名称");
+		assertThat(result).isEqualTo("前缀_工具名称");
+	}
+
+	@Test
+	void prefixedToolNameShouldSupportMixedChineseAndEnglish() {
+		String result = McpToolUtils.prefixedToolName("prefix前缀", "tool工具Name");
+		assertThat(result).isEqualTo("prefix前缀_tool工具Name");
+	}
+
+	@Test
+	void prefixedToolNameShouldRemoveSpecialCharactersButKeepChinese() {
+		String result = McpToolUtils.prefixedToolName("pre@fix前缀", "tool#工具$name");
+		assertThat(result).isEqualTo("prefix前缀_tool工具name");
+	}
+
+	@Test
+	void prefixedToolNameShouldHandleChineseWithHyphens() {
+		String result = McpToolUtils.prefixedToolName("前缀-test", "工具-name");
+		assertThat(result).isEqualTo("前缀_test_工具_name");
+	}
+
+	@Test
+	void prefixedToolNameShouldTruncateLongChineseStrings() {
+		// Create a string with Chinese characters that exceeds 64 characters
+		String longPrefix = "前缀".repeat(20); // 40 Chinese characters
+		String longToolName = "工具".repeat(20); // 40 Chinese characters
+		String result = McpToolUtils.prefixedToolName(longPrefix, longToolName);
+		assertThat(result).hasSize(64);
+		assertThat(result).endsWith("_" + "工具".repeat(20));
+	}
+
+	@Test
+	void prefixedToolNameShouldHandleChinesePunctuation() {
+		String result = McpToolUtils.prefixedToolName("前缀，测试", "工具。名称！");
+		assertThat(result).isEqualTo("前缀测试_工具名称");
+	}
+
+	@Test
+	void prefixedToolNameShouldHandleUnicodeBoundaries() {
+		// Test characters at the boundaries of the Chinese Unicode range
+		String result1 = McpToolUtils.prefixedToolName("prefix", "tool\u4e00"); // First
+																				// Chinese
+																				// character
+		assertThat(result1).isEqualTo("prefix_tool\u4e00");
+
+		String result2 = McpToolUtils.prefixedToolName("prefix", "tool\u9fa5"); // Last
+																				// Chinese
+																				// character
+		assertThat(result2).isEqualTo("prefix_tool\u9fa5");
+	}
+
+	@Test
+	void prefixedToolNameShouldExcludeNonChineseUnicodeCharacters() {
+		// Test with Japanese Hiragana (outside Chinese range)
+		String result1 = McpToolUtils.prefixedToolName("prefix", "toolあ"); // Japanese
+																			// Hiragana
+		assertThat(result1).isEqualTo("prefix_tool");
+
+		// Test with Korean characters (outside Chinese range)
+		String result2 = McpToolUtils.prefixedToolName("prefix", "tool한"); // Korean
+																			// character
+		assertThat(result2).isEqualTo("prefix_tool");
+
+		// Test with Arabic characters (outside Chinese range)
+		String result3 = McpToolUtils.prefixedToolName("prefix", "toolع"); // Arabic
+																			// character
+		assertThat(result3).isEqualTo("prefix_tool");
+	}
+
+	@Test
+	void prefixedToolNameShouldHandleEmojisAndSymbols() {
+		// Emojis and symbols should be removed
+		String result = McpToolUtils.prefixedToolName("prefix🚀", "tool工具😀name");
+		assertThat(result).isEqualTo("prefix_tool工具name");
+	}
+
+	@Test
+	void prefixedToolNameShouldPreserveNumbersWithChinese() {
+		String result = McpToolUtils.prefixedToolName("前缀123", "工具456名称");
+		assertThat(result).isEqualTo("前缀123_工具456名称");
+	}
+
+	@Test
+	void prefixedToolNameShouldSupportExtendedHanCharacters() {
+		// Test boundary character at end of CJK Unified Ideographs block
+		String result1 = McpToolUtils.prefixedToolName("prefix", "tool\u9fff"); // CJK
+																				// block
+																				// boundary
+		assertThat(result1).isEqualTo("prefix_tool\u9fff");
+
+		// Test CJK Extension A characters
+		String result2 = McpToolUtils.prefixedToolName("prefix", "tool\u3400"); // CJK Ext
+																				// A
+		assertThat(result2).isEqualTo("prefix_tool\u3400");
+	}
+
+	@Test
+	void prefixedToolNameShouldSupportCompatibilityIdeographs() {
+		// Test CJK Compatibility Ideographs
+		String result = McpToolUtils.prefixedToolName("prefix", "tool\uf900"); // Compatibility
+																				// ideograph
+		assertThat(result).isEqualTo("prefix_tool\uf900");
+	}
+
+	@Test
+	void prefixedToolNameShouldHandleAllHanScriptCharacters() {
+		// Mix of different Han character blocks: Extension A + CJK Unified +
+		// Compatibility
+		String result = McpToolUtils.prefixedToolName("前缀\u3400", "工具\u9fff名称\uf900");
+		assertThat(result).isEqualTo("前缀\u3400_工具\u9fff名称\uf900");
+	}
+
 	@Test
 	void constructorShouldBePrivate() throws Exception {
 		Constructor<McpToolUtils> constructor = McpToolUtils.class.getDeclaredConstructor();