feat: detect-bidi-characters rule (#95)

* Resolved conflicts in README

* Embedded anti-trojan-source and removed from dependencies

* Expanded README with an example

* Changed onCodePath with Program

* Improved code style as suggested in review

* Added rough location estimation in the error report

* feat: implement exact location for each bidi char

* Renamed to detect-bidi-characters and fixed first line offset in comments

* Added JSDoc in detectBidiCharacters

* Fixed MD034 - Bare URL used

* Apply suggestions from code review

Co-authored-by: Liran Tal <liran.tal@gmail.com>

Co-authored-by: Luciano Mammino <lucianomammino@gmail.com>
Co-authored-by: Liran Tal <liran.tal@gmail.com>

README.md

@@ -54,6 +54,7 @@ npm test
 
 | Name                                                                                         | Description                                                                                                                   | ⚠️  |
 | :------------------------------------------------------------------------------------------- | :---------------------------------------------------------------------------------------------------------------------------- | :-- |
+| [detect-bidi-characters](docs/rules/detect-bidi-characters.md)                               | Detects trojan source attacks that employ unicode bidi attacks to inject malicious code.                                      | ✅  |
 | [detect-buffer-noassert](docs/rules/detect-buffer-noassert.md)                               | Detects calls to "buffer" with "noAssert" flag set.                                                                           | ✅  |
 | [detect-child-process](docs/rules/detect-child-process.md)                                   | Detects instances of "child_process" & non-literal "exec()" calls.                                                            | ✅  |
 | [detect-disable-mustache-escape](docs/rules/detect-disable-mustache-escape.md)               | Detects "object.escapeMarkup = false", which can be used with some template engines to disable escaping of HTML entities.     | ✅  |

docs/rules/detect-bidi-characters.md

@@ -0,0 +1,50 @@
+# Detects trojan source attacks that employ unicode bidi attacks to inject malicious code (`security/detect-bidi-characters`)
+
+⚠️ This rule _warns_ in the ✅ `recommended` config.
+
+<!-- end auto-generated rule header -->
+
+Detects cases of [trojan source attacks](https://trojansource.codes) that employ unicode bidi attacks to inject malicious code
+
+## Why is Trojan Source important?
+
+The following publication on the topic of unicode characters attacks, dubbed [Trojan Source: Invisible Vulnerabilities](https://trojansource.codes/trojan-source.pdf), has caused a lot of concern from potential supply chain attacks where adversaries are able to inject malicious code into the source code of a project, slipping by unseen in the code review process.
+
+### An example
+
+As an example, take the following code where `RLO`, `LRI`, `PDI`, `IRI` are placeholders to visualise the respective dangerous unicode characters:
+
+```js
+#!/usr/bin/env node
+
+var accessLevel = "user";
+
+if (accessLevel != "userRLO LRI// Check if adminPDI IRI") {
+  console.log("You are an admin.");
+}
+```
+
+The code above, will be rendered by a text editor as follows:
+
+```js
+#!/usr/bin/env node
+
+var accessLevel = "user";
+
+if (accessLevel != "user") {
+  // Check if admin
+  console.log("You are an admin.");
+}
+```
+
+By looking at the rendered code above, a user reviewing this code might not notice the injected malicious unicode characters which are actually changing the semantic and the behaviour of the actual code.
+
+### More information
+
+For more information on the topic, you're welcome to read on the official website [trojansource.codes](https://trojansource.codes/) and the following [source code repository](https://github.com/nickboucher/trojan-source/) which contains the source code of the publication.
+
+### References
+
+- <https://certitude.consulting/blog/en/invisible-backdoor/>
+- <https://github.com/lirantal/anti-trojan-source/>
+- <https://github.com/lirantal/eslint-plugin-anti-trojan-source>

index.js

@@ -18,7 +18,8 @@ module.exports = {
     'detect-child-process': require('./rules/detect-child-process'),
     'detect-disable-mustache-escape': require('./rules/detect-disable-mustache-escape'),
     'detect-object-injection': require('./rules/detect-object-injection'),
-    'detect-new-buffer': require('./rules/detect-new-buffer')
+    'detect-new-buffer': require('./rules/detect-new-buffer'),
+    'detect-bidi-characters': require('./rules/detect-bidi-characters'),
   },
   rulesConfig: {
     'detect-unsafe-regex': 0,
@@ -33,7 +34,8 @@ module.exports = {
     'detect-child-process': 0,
     'detect-disable-mustache-escape': 0,
     'detect-object-injection': 0,
-    'detect-new-buffer': 0
+    'detect-new-buffer': 0,
+    'detect-bidi-characters': 0,
   },
   configs: {
     recommended: {
@@ -51,8 +53,9 @@ module.exports = {
         'security/detect-object-injection': 'warn',
         'security/detect-possible-timing-attacks': 'warn',
         'security/detect-pseudoRandomBytes': 'warn',
-        'security/detect-unsafe-regex': 'warn'
-      }
-    }
-  }
+        'security/detect-unsafe-regex': 'warn',
+        'security/detect-bidi-characters': 'warn',
+      },
+    },
+  },
 };

rules/detect-bidi-characters.js

@@ -0,0 +1,101 @@
+/**
+ * Detect trojan source attacks that employ unicode bidi attacks to inject malicious code
+ * @author Luciamo Mammino
+ * @author Simone Sanfratello
+ * @author Liran Tal
+ */
+
+'use strict';
+
+const dangerousBidiCharsRegexp = /[\u061C\u200E\u200F\u202A\u202B\u202C\u202D\u202E\u2066\u2067\u2068\u2069]/gu;
+
+/**
+ * Detects all the dangerous bidi characters in a given source text
+ *
+ * @param {object} options - Options
+ * @param {string} options.sourceText - The source text to search for dangerous bidi characters
+ * @param {number} options.firstLineOffset - The offset of the first line in the source text
+ * @returns {Array<{line: number, column: number}>} - An array of reports, each report is an
+ *    object with the line and column of the dangerous character
+ */
+function detectBidiCharacters({ sourceText, firstLineOffset }) {
+  const sourceTextToSearch = sourceText.toString();
+
+  const lines = sourceTextToSearch.split(/\r?\n/);
+
+  return lines.reduce((reports, line, lineIndex) => {
+    let match;
+    let offset = lineIndex == 0 ? firstLineOffset : 0;
+
+    while ((match = dangerousBidiCharsRegexp.exec(line)) !== null) {
+      reports.push({ line: lineIndex, column: offset + match.index });
+    }
+
+    return reports;
+  }, []);
+}
+
+function report({ context, node, tokens, message, firstLineOffset }) {
+  if (!tokens || !Array.isArray(tokens)) {
+    return;
+  }
+  tokens.forEach((token) => {
+    const reports = detectBidiCharacters({ sourceText: token.value, firstLineOffset: token.loc.start.column + firstLineOffset });
+
+    reports.forEach((report) => {
+      context.report({
+        node: node,
+        data: {
+          text: token.value,
+        },
+        loc: {
+          start: {
+            line: token.loc.start.line + report.line,
+            column: report.column,
+          },
+          end: {
+            line: token.loc.start.line + report.line,
+            column: report.column + 1,
+          },
+        },
+        message,
+      });
+    });
+  });
+}
+
+//------------------------------------------------------------------------------
+// Rule Definition
+//------------------------------------------------------------------------------
+
+module.exports = {
+  meta: {
+    type: 'error',
+    docs: {
+      description: 'Detects trojan source attacks that employ unicode bidi attacks to inject malicious code.',
+      category: 'Possible Security Vulnerability',
+      recommended: true,
+      url: 'https://github.com/eslint-community/eslint-plugin-security/blob/main/docs/rules/detect-bidi-characters.md',
+    },
+  },
+  create: function (context) {
+    return {
+      Program: function (node) {
+        report({
+          context,
+          node,
+          tokens: node.tokens,
+          firstLineOffset: 0,
+          message: "Detected potential trojan source attack with unicode bidi introduced in this code: '{{text}}'.",
+        });
+        report({
+          context,
+          node,
+          tokens: node.comments,
+          firstLineOffset: 2,
+          message: "Detected potential trojan source attack with unicode bidi introduced in this comment: '{{text}}'.",
+        });
+      },
+    };
+  },
+};

test/detect-bidi-characters.js

@@ -0,0 +1,74 @@
+'use strict';
+
+const RuleTester = require('eslint').RuleTester;
+const tester = new RuleTester();
+
+const ruleName = 'detect-bidi-characters';
+const Rule = require(`../rules/${ruleName}`);
+
+tester.run(ruleName, Rule, {
+  valid: [
+    {
+      code: `
+  var accessLevel = "user";
+  if (accessLevel != "user") { // Check if admin
+    console.log("You are an admin.");
+  }
+  `,
+    },
+  ],
+  invalid: [
+    {
+      code: `
+      var accessLevel = "user";
+      if (accessLevel != "user‮ ⁦// Check if admin⁩ ⁦") {
+          console.log("You are an admin.");
+      }
+      `,
+      errors: [
+        { message: /Detected potential trojan source attack with unicode bidi introduced in this code/i, line: 3, endLine: 3, column: 31, endColumn: 32 },
+        { message: /Detected potential trojan source attack with unicode bidi introduced in this code/i, line: 3, endLine: 3, column: 33, endColumn: 34 },
+        { message: /Detected potential trojan source attack with unicode bidi introduced in this code/i, line: 3, endLine: 3, column: 51, endColumn: 52 },
+        { message: /Detected potential trojan source attack with unicode bidi introduced in this code/i, line: 3, endLine: 3, column: 53, endColumn: 54 },
+      ],
+    },
+  ],
+});
+
+tester.run(`${ruleName} in comment-line`, Rule, {
+  valid: [
+    {
+      code: `
+  var isAdmin = false;
+  /* begin admins only */ if (isAdmin) {
+    console.log("You are an admin.");
+  /* end admins only */ }
+  `,
+    },
+  ],
+  invalid: [
+    {
+      code: `
+      var isAdmin = false;
+      /*‮ } ⁦if (isAdmin)⁩ ⁦ begin admins only */
+          console.log("You are an admin.");
+      /* end admins only ‮
+⁦*/
+      /* end admins only ‮ 
+ { ⁦*/
+        `,
+      errors: [
+        { message: /Detected potential trojan source attack with unicode bidi introduced in this comment/i, line: 3, endLine: 3, column: 9, endColumn: 10 },
+        { message: /Detected potential trojan source attack with unicode bidi introduced in this comment/i, line: 3, endLine: 3, column: 13, endColumn: 14 },
+        { message: /Detected potential trojan source attack with unicode bidi introduced in this comment/i, line: 3, endLine: 3, column: 26, endColumn: 27 },
+        { message: /Detected potential trojan source attack with unicode bidi introduced in this comment/i, line: 3, endLine: 3, column: 28, endColumn: 29 },
+
+        { message: /Detected potential trojan source attack with unicode bidi introduced in this comment/i, line: 5, endLine: 5, column: 26, endColumn: 27 },
+        { message: /Detected potential trojan source attack with unicode bidi introduced in this comment/i, line: 6, endLine: 6, column: 1, endColumn: 2 },
+
+        { message: /Detected potential trojan source attack with unicode bidi introduced in this comment/i, line: 7, endLine: 7, column: 26, endColumn: 27 },
+        { message: /Detected potential trojan source attack with unicode bidi introduced in this comment/i, line: 8, endLine: 8, column: 4, endColumn: 5 },
+      ],
+    },
+  ],
+});