ktg-plugin-marketplace/plugins/llm-security/tests/scanners/dep-token-overlap.test.mjs

// dep-token-overlap.test.mjs — B7 (v7.2.0) — typosquat token-overlap integration
//
// Verifies that dep-auditor's checkTyposquatting now flags packages with
// edit distance >= 3 that contain all tokens of a popular package plus a
// suspicious suffix (e.g. `lodash-utils` vs `lodash`).
//
// Builds an in-memory fixture in tmpdir per test instead of mutating the
// shared `tests/fixtures/dep-test/` fixture (which other tests assert
// exact finding counts against).

import { describe, it, before, after, beforeEach } from 'node:test';
import assert from 'node:assert/strict';
import { mkdtemp, writeFile, rm } from 'node:fs/promises';
import { tmpdir } from 'node:os';
import { join } from 'node:path';
import { resetCounter } from '../../scanners/lib/output.mjs';
import { scan } from '../../scanners/dep-auditor.mjs';

async function makeFixture(deps) {
  const dir = await mkdtemp(join(tmpdir(), 'llm-security-dep-token-'));
  await writeFile(
    join(dir, 'package.json'),
    JSON.stringify({ name: 'token-overlap-fixture', version: '1.0.0', dependencies: deps }, null, 2),
    'utf8',
  );
  return dir;
}

describe('dep-auditor B7 — token-overlap typosquat heuristic', () => {
  beforeEach(() => {
    resetCounter();
  });

  it('flags lodash-utils as token-overlap typosquat of lodash', async () => {
    const dir = await makeFixture({ 'lodash-utils': '^1.0.0' });
    try {
      const result = await scan(dir, { files: [] });
      const finding = result.findings.find(
        f => f.title.toLowerCase().includes('token-overlap') &&
             f.title.includes('lodash-utils') &&
             f.title.includes('lodash')
      );
      assert.ok(
        finding,
        `expected token-overlap finding for lodash-utils. Got: ${result.findings.map(f => f.title).join('; ')}`,
      );
      assert.equal(finding.severity, 'medium');
    } finally {
      await rm(dir, { recursive: true, force: true });
    }
  });

  it('flags react-helper as token-overlap typosquat of react', async () => {
    const dir = await makeFixture({ 'react-helper': '^1.0.0' });
    try {
      const result = await scan(dir, { files: [] });
      const finding = result.findings.find(
        f => f.title.toLowerCase().includes('token-overlap') &&
             f.title.includes('react-helper')
      );
      assert.ok(finding, `expected react-helper to be flagged. Got: ${result.findings.map(f => f.title).join('; ')}`);
    } finally {
      await rm(dir, { recursive: true, force: true });
    }
  });

  it('flags express-wrapper as token-overlap typosquat of express', async () => {
    const dir = await makeFixture({ 'express-wrapper': '^1.0.0' });
    try {
      const result = await scan(dir, { files: [] });
      const finding = result.findings.find(
        f => f.title.toLowerCase().includes('token-overlap') &&
             f.title.includes('express-wrapper')
      );
      assert.ok(finding, `expected express-wrapper to be flagged. Got: ${result.findings.map(f => f.title).join('; ')}`);
    } finally {
      await rm(dir, { recursive: true, force: true });
    }
  });

  it('does NOT flag legitimate package lodash (exact match)', async () => {
    const dir = await makeFixture({ 'lodash': '^4.17.0' });
    try {
      const result = await scan(dir, { files: [] });
      const typosquatFindings = result.findings.filter(
        f => f.title.toLowerCase().includes('typosquat')
      );
      assert.equal(typosquatFindings.length, 0, `lodash must not be flagged. Got: ${typosquatFindings.map(f => f.title).join('; ')}`);
    } finally {
      await rm(dir, { recursive: true, force: true });
    }
  });

  it('does NOT flag legitimate short-name tools from allowlist', async () => {
    // These are in knowledge/typosquat-allowlist.json (22 npm + 5 PyPI from v7.0.0).
    // The allowlist check fires BEFORE Levenshtein and BEFORE token-overlap.
    const dir = await makeFixture({
      'knip': '^5.0.0',
      'tsx': '^4.0.0',
      'nx': '^17.0.0',
      'rimraf': '^5.0.0',
    });
    try {
      const result = await scan(dir, { files: [] });
      const typosquatFindings = result.findings.filter(
        f => f.title.toLowerCase().includes('typosquat')
      );
      assert.equal(
        typosquatFindings.length, 0,
        `allowlisted short-name tools must not be flagged. Got: ${typosquatFindings.map(f => f.title).join('; ')}`,
      );
    } finally {
      await rm(dir, { recursive: true, force: true });
    }
  });

  it('does NOT flag packages with no suspicious suffix even with token overlap', async () => {
    // `react-router-dom` contains `react` token + `router` + `dom` — but neither
    // `router` nor `dom` is in TYPOSQUAT_SUSPICIOUS_TOKENS. This is a legitimate
    // ecosystem name and must pass.
    const dir = await makeFixture({ 'react-router-dom': '^6.0.0' });
    try {
      const result = await scan(dir, { files: [] });
      const tokenOverlapFindings = result.findings.filter(
        f => f.title.toLowerCase().includes('token-overlap')
      );
      assert.equal(
        tokenOverlapFindings.length, 0,
        `react-router-dom must not be flagged as token-overlap typosquat`,
      );
    } finally {
      await rm(dir, { recursive: true, force: true });
    }
  });

  it('does NOT flag packages whose all tokens match a popular package (subset, not superset)', async () => {
    // `react` itself has tokens {react} which equal the popular `react` token set.
    // The heuristic requires declared ⊃ popular (strict superset), not equal.
    const dir = await makeFixture({ 'react': '^18.0.0' });
    try {
      const result = await scan(dir, { files: [] });
      const tokenOverlapFindings = result.findings.filter(
        f => f.title.toLowerCase().includes('token-overlap')
      );
      assert.equal(tokenOverlapFindings.length, 0);
    } finally {
      await rm(dir, { recursive: true, force: true });
    }
  });
});