sanitize nosql

Signed-off-by: ale <ale@manalejandro.com>

app/api/search/route.ts

@@ -11,13 +11,101 @@ interface HashDocument {
   created_at?: string;
 }
 
+// Maximum allowed query length
+const MAX_QUERY_LENGTH = 1000;
+
+// Characters that could be used in NoSQL/Elasticsearch injection attacks
+const DANGEROUS_PATTERNS = [
+  /[{}\[\]]/g,           // JSON structure characters
+  /\$[a-zA-Z]/g,         // MongoDB-style operators
+  /\\u[0-9a-fA-F]{4}/g,  // Unicode escapes
+  /<script/gi,           // XSS attempts
+  /javascript:/gi,       // XSS attempts
+];
+
+/**
+ * Sanitize input to prevent NoSQL injection attacks
+ * For hash lookups, we only need alphanumeric characters and $
+ * For plaintext, we allow more characters but sanitize dangerous patterns
+ */
+function sanitizeInput(input: string): string {
+  // Trim and take first word only
+  let sanitized = input.trim().split(/\s+/)[0] || '';
+  
+  // Limit length
+  if (sanitized.length > MAX_QUERY_LENGTH) {
+    sanitized = sanitized.substring(0, MAX_QUERY_LENGTH);
+  }
+  
+  // Remove null bytes
+  sanitized = sanitized.replace(/\0/g, '');
+  
+  // Check for dangerous patterns
+  for (const pattern of DANGEROUS_PATTERNS) {
+    sanitized = sanitized.replace(pattern, '');
+  }
+  
+  return sanitized;
+}
+
+/**
+ * Validate that the input is safe for use in Elasticsearch queries
+ */
+function isValidInput(input: string): boolean {
+  // Check for empty input
+  if (!input || input.length === 0) {
+    return false;
+  }
+  
+  // Check for excessively long input
+  if (input.length > MAX_QUERY_LENGTH) {
+    return false;
+  }
+  
+  // Check for control characters (except normal whitespace)
+  if (/[\x00-\x08\x0B\x0C\x0E-\x1F\x7F]/.test(input)) {
+    return false;
+  }
+  
+  return true;
+}
+
 export async function POST(request: NextRequest) {
   try {
-    const { query } = await request.json();
+    const body = await request.json();
+    
+    // Validate request body structure
+    if (!body || typeof body !== 'object') {
+      return NextResponse.json(
+        { error: 'Invalid request body' },
+        { status: 400 }
+      );
+    }
 
+    const { query } = body;
+
+    // Validate query type
     if (!query || typeof query !== 'string') {
       return NextResponse.json(
-        { error: 'Query parameter is required' },
+        { error: 'Query parameter is required and must be a string' },
+        { status: 400 }
+      );
+    }
+
+    // Validate input before processing
+    if (!isValidInput(query)) {
+      return NextResponse.json(
+        { error: 'Invalid query: contains forbidden characters or is too long' },
+        { status: 400 }
+      );
+    }
+
+    // Sanitize input
+    const cleanQuery = sanitizeInput(query);
+    
+    if (!cleanQuery) {
+      return NextResponse.json(
+        { error: 'Invalid query: only whitespace or invalid characters provided' },
         { status: 400 }
       );
     }
@@ -25,15 +113,6 @@ export async function POST(request: NextRequest) {
     // Ensure index exists
     await initializeIndex();
 
-    const cleanQuery = query.trim().split(/\s+/)[0];
-    
-    if (!cleanQuery) {
-      return NextResponse.json(
-        { error: 'Invalid query: only whitespace provided' },
-        { status: 400 }
-      );
-    }
-
     const cleanQueryLower = cleanQuery.toLowerCase();
     const hashType = detectHashType(cleanQueryLower);