fix(knowledge): reject non-alphanumeric file extensions from document names (#3816)

* fix(knowledge): reject non-alphanumeric file extensions from document names

* fix(knowledge): improve error message when extension is non-alphanumeric

Author: waleedlatif1

apps/sim/lib/knowledge/documents/parser-extension.ts

@@ -29,9 +29,8 @@ export function resolveParserExtension(
   mimeType?: string,
   fallback?: string
 ): string {
-  const filenameExtension = filename.includes('.')
-    ? filename.split('.').pop()?.toLowerCase()
-    : undefined
+  const raw = filename.includes('.') ? filename.split('.').pop()?.toLowerCase() : undefined
+  const filenameExtension = raw && /^[a-z0-9]+$/.test(raw) ? raw : undefined
 
   if (filenameExtension && isSupportedParserExtension(filenameExtension)) {
     return filenameExtension

apps/sim/lib/uploads/utils/validation.ts

@@ -137,12 +137,13 @@ export interface FileValidationError {
  * Validate if a file type is supported for document processing
  */
 export function validateFileType(fileName: string, mimeType: string): FileValidationError | null {
-  const extension = path.extname(fileName).toLowerCase().substring(1) as SupportedDocumentExtension
+  const raw = path.extname(fileName).toLowerCase().substring(1)
+  const extension = (/^[a-z0-9]+$/.test(raw) ? raw : '') as SupportedDocumentExtension
 
   if (!SUPPORTED_DOCUMENT_EXTENSIONS.includes(extension)) {
     return {
       code: 'UNSUPPORTED_FILE_TYPE',
-      message: `Unsupported file type: ${extension}. Supported types are: ${SUPPORTED_DOCUMENT_EXTENSIONS.join(', ')}`,
+      message: `Unsupported file type${extension ? `: ${extension}` : ` for "${fileName}"`}. Supported types are: ${SUPPORTED_DOCUMENT_EXTENSIONS.join(', ')}`,
       supportedTypes: [...SUPPORTED_DOCUMENT_EXTENSIONS],
     }
   }
@@ -221,15 +222,16 @@ export function validateMediaFileType(
   fileName: string,
   mimeType: string
 ): FileValidationError | null {
-  const extension = path.extname(fileName).toLowerCase().substring(1)
+  const raw = path.extname(fileName).toLowerCase().substring(1)
+  const extension = /^[a-z0-9]+$/.test(raw) ? raw : ''
 
   const isAudio = SUPPORTED_AUDIO_EXTENSIONS.includes(extension as SupportedAudioExtension)
   const isVideo = SUPPORTED_VIDEO_EXTENSIONS.includes(extension as SupportedVideoExtension)
 
   if (!isAudio && !isVideo) {
     return {
       code: 'UNSUPPORTED_FILE_TYPE',
-      message: `Unsupported media file type: ${extension}. Supported audio types: ${SUPPORTED_AUDIO_EXTENSIONS.join(', ')}. Supported video types: ${SUPPORTED_VIDEO_EXTENSIONS.join(', ')}`,
+      message: `Unsupported media file type${extension ? `: ${extension}` : ` for "${fileName}"`}. Supported audio types: ${SUPPORTED_AUDIO_EXTENSIONS.join(', ')}. Supported video types: ${SUPPORTED_VIDEO_EXTENSIONS.join(', ')}`,
       supportedTypes: [...SUPPORTED_AUDIO_EXTENSIONS, ...SUPPORTED_VIDEO_EXTENSIONS],
     }
   }