- Extension package: Manifest V3, background service worker, content scripts - Phishing detection engine with heuristic analysis (typosquatting, entropy, TLD, brand impersonation) - Local URL caching layer (Storage API) for <100ms cached lookups - Popup UI with protection status, stats, and phishing report button - Options page for settings management (blocked/allowed domains, feature toggles) - Server-side extension routes: URL check, phishing report, auth, stats, exposure check - Tier-aware feature gating (Basic/Plus/Premium) - 25 passing tests for phishing detection heuristics - Declarative net request rules for known phishing patterns - DarkWatch integration for credential exposure checks - Firefox compatibility layer via build modes Co-Authored-By: Paperclip <noreply@paperclip.ing>
278 lines
8.8 KiB
TypeScript
278 lines
8.8 KiB
TypeScript
import { ThreatType, UrlVerdict, ThreatInfo } from '../types';
|
|
|
|
export class PhishingDetector {
|
|
private knownSuspiciousTlds: Set<string> = new Set([
|
|
'.tk', '.ml', '.ga', '.cf', '.gq', '.xyz', '.top', '.click', '.link', '.work',
|
|
]);
|
|
|
|
private commonBrands: Map<string, string[]> = new Map([
|
|
['google', ['gmail', 'drive', 'docs', 'maps', 'play', 'chrome', 'youtube']],
|
|
['apple', ['icloud', 'appstore', 'icloud_content', 'appleid']],
|
|
['amazon', ['aws', 'amazonaws', 'amazon-adsystem', 'prime-video']],
|
|
['microsoft', ['office', 'outlook', 'onedrive', 'teams', 'azure', 'windows']],
|
|
['facebook', ['fb', 'fbcdn', 'instagram', 'whatsapp', 'messenger']],
|
|
['paypal', ['paypalobjects', 'paypal-web', 'xoom']],
|
|
['netflix', ['nflximg', 'nflxso', 'nflxvideo', 'nflxext']],
|
|
['bank', ['chase', 'bofa', 'wellsfargo', 'citi', 'hsbc', 'barclays']],
|
|
]);
|
|
|
|
analyzeUrl(url: string): { verdict: UrlVerdict; threats: ThreatInfo[]; score: number } {
|
|
const threats: ThreatInfo[] = [];
|
|
let score = 0;
|
|
|
|
try {
|
|
const parsed = new URL(url);
|
|
const hostname = parsed.hostname.toLowerCase();
|
|
const domainParts = hostname.split('.');
|
|
const tld = domainParts[domainParts.length - 1];
|
|
|
|
score += this.checkTld(tld, domainParts, threats);
|
|
score += this.checkEntropy(parsed.pathname + parsed.search, threats);
|
|
score += this.checkTyposquatting(hostname, threats);
|
|
score += this.checkIpAddress(hostname, threats);
|
|
score += this.checkLongUrl(url, threats);
|
|
score += this.checkSubdomainDepth(domainParts, threats);
|
|
score += this.checkHttpsProtocol(parsed.protocol, threats);
|
|
score += this.checkRedirectPatterns(parsed.search, threats);
|
|
score += this.checkEncodedChars(url, threats);
|
|
score += this.checkBrandImpersonation(hostname, threats);
|
|
} catch {
|
|
return {
|
|
verdict: UrlVerdict.UNKNOWN,
|
|
threats: [{ type: ThreatType.PHISHING_HEURISTIC, severity: 3, source: 'heuristic', description: 'Malformed URL' }],
|
|
score: 30,
|
|
};
|
|
}
|
|
|
|
const verdict = score >= 70 ? UrlVerdict.PHISHING
|
|
: score >= 40 ? UrlVerdict.SUSPICIOUS
|
|
: score >= 20 ? UrlVerdict.SPAM
|
|
: UrlVerdict.SAFE;
|
|
|
|
return { verdict, threats, score };
|
|
}
|
|
|
|
private checkTld(tld: string, parts: string[], threats: ThreatInfo[]): number {
|
|
if (this.knownSuspiciousTlds.has(`.${tld}`)) {
|
|
threats.push({
|
|
type: ThreatType.DOMAIN_AGE,
|
|
severity: 4,
|
|
source: 'heuristic',
|
|
description: `Suspicious TLD: .${tld}`,
|
|
});
|
|
return 25;
|
|
}
|
|
if (parts.length === 1) {
|
|
threats.push({
|
|
type: ThreatType.DOMAIN_AGE,
|
|
severity: 3,
|
|
source: 'heuristic',
|
|
description: 'Single-label domain (possible local or new domain)',
|
|
});
|
|
return 15;
|
|
}
|
|
return 0;
|
|
}
|
|
|
|
private checkEntropy(pathname: string, threats: ThreatInfo[]): number {
|
|
if (!pathname || pathname.length < 20) return 0;
|
|
const entropy = this.calculateEntropy(pathname);
|
|
if (entropy > 4.5) {
|
|
threats.push({
|
|
type: ThreatType.URL_ENTROPY,
|
|
severity: 4,
|
|
source: 'heuristic',
|
|
description: `High URL path entropy (${entropy.toFixed(2)}) suggests obfuscation`,
|
|
});
|
|
return 20;
|
|
}
|
|
return 0;
|
|
}
|
|
|
|
private checkTyposquatting(hostname: string, threats: ThreatInfo[]): number {
|
|
for (const [brand, subdomains] of this.commonBrands) {
|
|
const brandParts = hostname.split('.');
|
|
const mainDomain = brandParts.slice(0, -1).join('.');
|
|
const firstLabel = mainDomain.split('.')[0];
|
|
|
|
if (mainDomain.includes(brand) && mainDomain !== brand) {
|
|
const editDist = this.levenshteinDistance(firstLabel, brand);
|
|
if (editDist <= 2 && editDist > 0) {
|
|
threats.push({
|
|
type: ThreatType.TYPOSQUAT,
|
|
severity: 5,
|
|
source: 'heuristic',
|
|
description: `Possible typosquat of "${brand}" (edit distance: ${editDist})`,
|
|
});
|
|
return 35;
|
|
}
|
|
}
|
|
|
|
const editDist = this.levenshteinDistance(firstLabel, brand);
|
|
if (editDist <= 2 && editDist > 0 && firstLabel.length >= brand.length - 1) {
|
|
threats.push({
|
|
type: ThreatType.TYPOSQUAT,
|
|
severity: 5,
|
|
source: 'heuristic',
|
|
description: `Possible typosquat of "${brand}" (edit distance: ${editDist})`,
|
|
});
|
|
return 35;
|
|
}
|
|
|
|
for (const sub of subdomains) {
|
|
if (hostname.includes(sub) && !hostname.startsWith(`${sub}.` + brandParts[brandParts.length - 1])) {
|
|
threats.push({
|
|
type: ThreatType.TYPOSQUAT,
|
|
severity: 3,
|
|
source: 'heuristic',
|
|
description: `Domain contains "${sub}" but not an official ${brand} subdomain`,
|
|
});
|
|
return 15;
|
|
}
|
|
}
|
|
}
|
|
return 0;
|
|
}
|
|
|
|
private checkIpAddress(hostname: string, threats: ThreatInfo[]): number {
|
|
const ipPattern = /^\d{1,3}\.\d{1,3}\.\d{1,3}\.\d{1,3}$/;
|
|
if (ipPattern.test(hostname) && hostname !== '127.0.0.1' && hostname !== 'localhost') {
|
|
threats.push({
|
|
type: ThreatType.PHISHING_HEURISTIC,
|
|
severity: 4,
|
|
source: 'heuristic',
|
|
description: `IP address used as hostname: ${hostname}`,
|
|
});
|
|
return 25;
|
|
}
|
|
return 0;
|
|
}
|
|
|
|
private checkLongUrl(url: string, threats: ThreatInfo[]): number {
|
|
if (url.length > 200) {
|
|
threats.push({
|
|
type: ThreatType.PHISHING_HEURISTIC,
|
|
severity: 3,
|
|
source: 'heuristic',
|
|
description: `Unusually long URL (${url.length} chars)`,
|
|
});
|
|
return 15;
|
|
}
|
|
return 0;
|
|
}
|
|
|
|
private checkSubdomainDepth(parts: string[], threats: ThreatInfo[]): number {
|
|
if (parts.length > 5) {
|
|
threats.push({
|
|
type: ThreatType.PHISHING_HEURISTIC,
|
|
severity: 3,
|
|
source: 'heuristic',
|
|
description: `Deep subdomain nesting (${parts.length} levels)`,
|
|
});
|
|
return 15;
|
|
}
|
|
return 0;
|
|
}
|
|
|
|
private checkHttpsProtocol(protocol: string, threats: ThreatInfo[]): number {
|
|
if (protocol === 'http:') {
|
|
threats.push({
|
|
type: ThreatType.MIXED_CONTENT,
|
|
severity: 2,
|
|
source: 'heuristic',
|
|
description: 'Page loaded over HTTP (not HTTPS)',
|
|
});
|
|
return 10;
|
|
}
|
|
return 0;
|
|
}
|
|
|
|
private checkRedirectPatterns(query: string, threats: ThreatInfo[]): number {
|
|
const redirectParams = ['redirect', 'url', 'dest', 'return', 'next', 'target'];
|
|
let count = 0;
|
|
for (const param of redirectParams) {
|
|
if (query.includes(`${param}=`)) count++;
|
|
}
|
|
if (count >= 2) {
|
|
threats.push({
|
|
type: ThreatType.REDIRECT_CHAIN,
|
|
severity: 3,
|
|
source: 'heuristic',
|
|
description: `Multiple redirect parameters detected (${count})`,
|
|
});
|
|
return 15;
|
|
}
|
|
return 0;
|
|
}
|
|
|
|
private checkEncodedChars(url: string, threats: ThreatInfo[]): number {
|
|
const encodedPattern = /(%[0-9a-fA-F]{2}){3,}/g;
|
|
const matches = url.match(encodedPattern);
|
|
if (matches && matches.length > 0) {
|
|
threats.push({
|
|
type: ThreatType.URL_ENTROPY,
|
|
severity: 3,
|
|
source: 'heuristic',
|
|
description: 'Excessive URL encoding detected',
|
|
});
|
|
return 15;
|
|
}
|
|
return 0;
|
|
}
|
|
|
|
private checkBrandImpersonation(hostname: string, threats: ThreatInfo[]): number {
|
|
const impersonationPatterns = [
|
|
/login[-_]?(secure|portal|page|form)/i,
|
|
/account[-_]?(verify|confirm|update)/i,
|
|
/secure[-_]?(signin|auth|login)/i,
|
|
/wallet[-_]?(connect|link|verify)/i,
|
|
];
|
|
for (const pattern of impersonationPatterns) {
|
|
if (pattern.test(hostname)) {
|
|
threats.push({
|
|
type: ThreatType.PHISHING_HEURISTIC,
|
|
severity: 4,
|
|
source: 'heuristic',
|
|
description: `Common phishing pattern detected: ${hostname}`,
|
|
});
|
|
return 20;
|
|
}
|
|
}
|
|
return 0;
|
|
}
|
|
|
|
private calculateEntropy(str: string): number {
|
|
const freq: Record<string, number> = {};
|
|
for (const char of str) {
|
|
freq[char] = (freq[char] || 0) + 1;
|
|
}
|
|
let entropy = 0;
|
|
const len = str.length;
|
|
for (const count of Object.values(freq)) {
|
|
const p = count / len;
|
|
entropy -= p * Math.log2(p);
|
|
}
|
|
return entropy;
|
|
}
|
|
|
|
private levenshteinDistance(a: string, b: string): number {
|
|
const matrix: number[][] = [];
|
|
for (let i = 0; i <= b.length; i++) matrix[i] = [i];
|
|
for (let j = 0; j <= a.length; j++) matrix[0][j] = j;
|
|
for (let i = 1; i <= b.length; i++) {
|
|
for (let j = 1; j <= a.length; j++) {
|
|
matrix[i][j] = b[i - 1] === a[j - 1]
|
|
? matrix[i - 1][j - 1]
|
|
: Math.min(
|
|
matrix[i - 1][j - 1] + 1,
|
|
matrix[i][j - 1] + 1,
|
|
matrix[i - 1][j] + 1,
|
|
);
|
|
}
|
|
}
|
|
return matrix[b.length][a.length];
|
|
}
|
|
}
|
|
|
|
export const phishingDetector = new PhishingDetector();
|