fix: Remove non-compliant characters from MCP tool names for API compatibility (#11903)

* Initial plan

* fix: address PR review feedback - remove dots/colons, add error handling and logging

Co-authored-by: DeJeune <67425183+DeJeune@users.noreply.github.com>

* refactor: improve error handling consistency and default values

Co-authored-by: DeJeune <67425183+DeJeune@users.noreply.github.com>

---------

Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com>
Co-authored-by: DeJeune <67425183+DeJeune@users.noreply.github.com>

src/main/utils/__tests__/mcp.test.ts

@@ -202,7 +202,7 @@ describe('buildFunctionCallToolName', () => {
       expect(result).not.toMatch(/[\u4e00-\u9fff]/) // No Chinese characters
       expect(result).toContain('ocr') // OCR is lowercased
       // Should only contain ASCII characters (lowercase)
-      expect(result).toMatch(/^[a-z_][a-z0-9_.\-:]*$/)
+      expect(result).toMatch(/^[a-z_][a-z0-9_-]*$/)
     })
 
     it('should distinguish between different Chinese OCR tools', () => {
@@ -219,7 +219,7 @@ describe('buildFunctionCallToolName', () => {
 
       // All should be ASCII-only valid tool names
       tools.forEach((tool) => {
-        expect(tool).toMatch(/^[a-zA-Z_][a-zA-Z0-9_.\-:]*$/)
+        expect(tool).toMatch(/^[a-z_][a-z0-9_-]*$/)
         expect(tool).not.toMatch(/[\u4e00-\u9fff]/) // No Chinese characters
       })
 
@@ -234,7 +234,7 @@ describe('buildFunctionCallToolName', () => {
     it('should handle Japanese characters with Romaji transliteration', () => {
       const result = buildFunctionCallToolName('server', 'ユーザー検索')
       // Should be ASCII-only (Japanese characters are transliterated to Romaji)
-      expect(result).toMatch(/^[a-zA-Z_][a-zA-Z0-9_.\-:]*$/)
+      expect(result).toMatch(/^[a-z_][a-z0-9_-]*$/)
       // Should not contain original Japanese characters
       expect(result).not.toMatch(/[\u3040-\u309f\u30a0-\u30ff]/)
     })
@@ -242,7 +242,7 @@ describe('buildFunctionCallToolName', () => {
     it('should handle Korean characters with romanization', () => {
       const result = buildFunctionCallToolName('server', '사용자검색')
       // Should be ASCII-only
-      expect(result).toMatch(/^[a-zA-Z_][a-zA-Z0-9_.\-:]*$/)
+      expect(result).toMatch(/^[a-z_][a-z0-9_-]*$/)
       // Should not contain original Korean characters
       expect(result).not.toMatch(/[\uac00-\ud7af]/)
     })
@@ -256,7 +256,7 @@ describe('buildFunctionCallToolName', () => {
       expect(result).toContain('yong_hu')
       expect(result).toContain('ming_cheng')
       // Final result should be ASCII-only (lowercase)
-      expect(result).toMatch(/^[a-z_][a-z0-9_.\-:]*$/)
+      expect(result).toMatch(/^[a-z_][a-z0-9_-]*$/)
     })
 
     it('should transliterate Chinese and replace special symbols', () => {
@@ -270,7 +270,7 @@ describe('buildFunctionCallToolName', () => {
       expect(result).toContain('shang_chuan')
       expect(result).toContain('gong_ju')
       // Should be ASCII-only (lowercase)
-      expect(result).toMatch(/^[a-z_][a-z0-9_.\-:]*$/)
+      expect(result).toMatch(/^[a-z_][a-z0-9_-]*$/)
     })
 
     it('should produce AI model compatible tool names', () => {
@@ -279,9 +279,9 @@ describe('buildFunctionCallToolName', () => {
       testCases.forEach((testCase) => {
         const result = buildFunctionCallToolName('server', testCase)
         // Must start with letter or underscore
-        expect(result).toMatch(/^[a-zA-Z_]/)
-        // Must only contain a-z, A-Z, 0-9, _, -, ., :
-        expect(result).toMatch(/^[a-zA-Z0-9_.\-:]+$/)
+        expect(result).toMatch(/^[a-z_]/)
+        // Must only contain a-z, 0-9, _, -
+        expect(result).toMatch(/^[a-z0-9_-]+$/)
         // Must be <= 64 characters
         expect(result.length).toBeLessThanOrEqual(64)
       })

src/main/utils/mcp.ts

@@ -1,5 +1,8 @@
+import { loggerService } from '@logger'
 import { transliterate } from 'transliteration'
 
+const logger = loggerService.withContext('Utils:MCP')
+
 /**
  * Transliterates non-ASCII text (including CJK characters) to ASCII-compatible format.
  *
@@ -8,31 +11,58 @@ import { transliterate } from 'transliteration'
  *
  * @param text - The input string to transliterate, may contain Unicode characters including CJK
  * @returns A lowercase ASCII string with spaces converted to underscores and special characters removed,
- *          preserving only alphanumeric characters, underscores, dots, hyphens, and colons
+ *          preserving only alphanumeric characters, underscores, and hyphens
  *
  * @example
  * ```typescript
  * transliterateToAscii("Hello World") // returns "hello_world"
  * transliterateToAscii("你好世界") // returns transliterated version with underscores
- * transliterateToAscii("Café-123") // returns "cafe_123"
+ * transliterateToAscii("Café-123") // returns "cafe-123"
  * ```
  */
 function transliterateToAscii(text: string): string {
-  // Use transliteration library which supports CJK (Chinese, Japanese, Korean)
-  const result = transliterate(text, {
-    // Unknown/special characters become underscores
-    unknown: '_',
-    ignore: []
-  })
+  // Input validation
+  if (!text || typeof text !== 'string') {
+    logger.warn('Invalid input to transliterateToAscii', { text })
+    return 'invalid_input'
+  }
 
-  // Convert to lowercase, remove spaces, and clean up special chars
-  return result
-    .toLowerCase()
-    .replace(/\s+/g, '_')
-    .replace(/[^a-z0-9_.\-:]/g, '_')
+  try {
+    // Use transliteration library which supports CJK (Chinese, Japanese, Korean)
+    const result = transliterate(text, {
+      // Unknown/special characters become underscores
+      unknown: '_',
+      ignore: []
+    })
+
+    logger.debug('Transliteration successful', { input: text, output: result })
+
+    // Convert to lowercase, remove spaces, and clean up special chars
+    // Only preserve a-z, 0-9, underscores, and hyphens (OpenAI/Anthropic API compatible)
+    return result
+      .toLowerCase()
+      .replace(/\s+/g, '_')
+      .replace(/[^a-z0-9_-]/g, '_')
+  } catch (error) {
+    logger.error('Transliteration failed, falling back to ASCII-only mode', { text, error })
+    // Fallback: keep only ASCII alphanumeric, underscores, and hyphens for consistency
+    return text
+      .toLowerCase()
+      .replace(/[^a-z0-9_-]/g, '_')
+  }
 }
 
 export function buildFunctionCallToolName(serverName: string, toolName: string, serverId?: string) {
+  // Input validation with descriptive fallbacks to indicate invalid input
+  if (!serverName || typeof serverName !== 'string') {
+    logger.warn('Invalid serverName provided', { serverName })
+    serverName = 'invalid_server'
+  }
+  if (!toolName || typeof toolName !== 'string') {
+    logger.warn('Invalid toolName provided', { toolName })
+    toolName = 'invalid_tool'
+  }
+
   // First, transliterate non-ASCII characters to ASCII
   const transliteratedServer = transliterateToAscii(serverName.trim())
   const transliteratedTool = transliterateToAscii(toolName.trim())
@@ -65,11 +95,11 @@ export function buildFunctionCallToolName(serverName: string, toolName: string,
   }
 
   // Replace invalid characters with underscores
-  // Keep only a-z, 0-9, underscores, dashes, dots, colons (AI model compatible)
-  name = name.replace(/[^a-z0-9_.\-:]/g, '_')
+  // Keep only a-z, 0-9, underscores, dashes (OpenAI/Anthropic API compatible)
+  name = name.replace(/[^a-z0-9_-]/g, '_')
 
   // Ensure name starts with a letter or underscore (AI model requirement)
-  if (!/^[a-zA-Z_]/.test(name)) {
+  if (!/^[a-z_]/.test(name)) {
     name = `tool_${name}`
   }