Revised implementation

CHANGELOG.md

@@ -1,5 +1,9 @@
 # CHANGELOG
 
+## UNRELEASED
+
+- Update to handle back-slashes
+
 ## 7.0.4
 
 - Updates get-func-name to 2.0.2

src/__tests__/index.test.ts

@@ -16,7 +16,7 @@ describe("sanitizeUrl", () => {
   });
 
   it("does not alter https URLs with alphanumeric characters", () => {
-    expect(sanitizeUrl("https://example.com")).toBe("https://example.com");
+    expect(sanitizeUrl("https://example.com")).toBe("https://example.com/");
   });
 
   it("does not alter https URLs with ports with alphanumeric characters", () => {
@@ -147,6 +147,28 @@ describe("sanitizeUrl", () => {
     });
   });
 
+  it("backslash prefixed attack vectors", () => {
+    const attackVectors = [
+      "\fjavascript:alert()",
+      "\vjavascript:alert()",
+      "\tjavascript:alert()",
+      "\njavascript:alert()",
+      "\rjavascript:alert()",
+      "\u0000javascript:alert()",
+      "\u0001javascript:alert()",
+    ];
+
+    attackVectors.forEach((vector) => {
+      expect(sanitizeUrl(vector)).toBe(BLANK_URL);
+    });
+  });
+
+  it("reverses backslashes", () => {
+    const attack = "\\j\\av\\a\\s\\cript:alert()";
+
+    expect(sanitizeUrl(attack)).toBe("/j/av/a/s/cript:alert()");
+  });
+
   describe("invalid protocols", () => {
     describe.each(["javascript", "data", "vbscript"])("%s", (protocol) => {
       it(`replaces ${protocol} urls with ${BLANK_URL}`, () => {

src/index.ts

@@ -5,22 +5,25 @@ import {
   htmlEntitiesRegex,
   invalidProtocolRegex,
   relativeFirstCharacters,
-  urlSchemeRegex,
   whitespaceEscapeCharsRegex,
+  urlSchemeRegex,
 } from "./constants";
 
 function isRelativeUrlWithoutProtocol(url: string): boolean {
   return relativeFirstCharacters.indexOf(url[0]) > -1;
 }
 
-// adapted from https://stackoverflow.com/a/29824550/2601552
 function decodeHtmlCharacters(str: string) {
   const removedNullByte = str.replace(ctrlCharactersRegex, "");
   return removedNullByte.replace(htmlEntitiesRegex, (match, dec) => {
     return String.fromCharCode(dec);
   });
 }
 
+function isValidUrl(url: string): boolean {
+  return URL.canParse(url);
+}
+
 function decodeURI(uri: string): string {
   try {
     return decodeURIComponent(uri);
@@ -36,8 +39,9 @@ export function sanitizeUrl(url?: string): string {
   if (!url) {
     return BLANK_URL;
   }
+
   let charsToDecode;
-  let decodedUrl = decodeURI(url);
+  let decodedUrl = decodeURI(url.trim());
 
   do {
     decodedUrl = decodeHtmlCharacters(decodedUrl)
@@ -54,7 +58,9 @@ export function sanitizeUrl(url?: string): string {
       decodedUrl.match(htmlCtrlEntityRegex) ||
       decodedUrl.match(whitespaceEscapeCharsRegex);
   } while (charsToDecode && charsToDecode.length > 0);
+
   const sanitizedUrl = decodedUrl;
+
   if (!sanitizedUrl) {
     return BLANK_URL;
   }
@@ -63,17 +69,39 @@ export function sanitizeUrl(url?: string): string {
     return sanitizedUrl;
   }
 
-  const urlSchemeParseResults = sanitizedUrl.match(urlSchemeRegex);
+  // Remove any leading whitespace before checking the URL scheme
+  const trimmedUrl = sanitizedUrl.trimStart();
+  const urlSchemeParseResults = trimmedUrl.match(urlSchemeRegex);
 
   if (!urlSchemeParseResults) {
     return sanitizedUrl;
   }
 
-  const urlScheme = urlSchemeParseResults[0];
+  const urlScheme = urlSchemeParseResults[0].toLowerCase().trim();
 
   if (invalidProtocolRegex.test(urlScheme)) {
     return BLANK_URL;
   }
 
-  return sanitizedUrl;
+  const backSanitized = trimmedUrl.replace(/\\/g, "/");
+
+  // Handle special cases for mailto: and custom deep-link protocols
+  if (urlScheme === "mailto:" || urlScheme.includes("://")) {
+    return backSanitized;
+  }
+
+  // For http and https URLs, perform additional validation
+  if (urlScheme === "http:" || urlScheme === "https:") {
+    if (!isValidUrl(backSanitized)) {
+      return BLANK_URL;
+    }
+
+    const url = new URL(backSanitized);
+    url.protocol = url.protocol.toLowerCase();
+    url.hostname = url.hostname.toLowerCase();
+
+    return url.toString();
+  }
+
+  return backSanitized;
 }