How to Implement a Secure Pattern Lock for Mobile Apps

Pattern lock — an Android original, less common on iOS (no system equivalent). It works like a PIN: local device unlock without sending the secret to the server. But the pattern has specific security issues you must understand before implementation. Our experience with mobile authentication (over 20 projects) shows: a proper implementation with the right crypto stack makes the pattern convenient and secure enough for most scenarios.

Security Considerations

Why is Pattern Lock Vulnerable to Shoulder Surfing?

Finger smudges remain on the screen. Research by Abraham and colleagues (2010) showed that most users draw L-, Z-, or S-shaped patterns — the 12 most popular patterns cover ~20% of the user base. A 9-dot pattern can be visually reconstructed from 1–2 meters under the correct lighting angle. In our practice, we use a line-hiding timer and fade animation to minimize leakage.

How is the Pattern Encoded and Cryptographically Protected?

Standard 3×3 grid — 9 dots indexed 0–8. The pattern is a sequence of indices. Minimum length is 4 dots (Android requirement). From 9 dots with minimum length 4, about 389,112 possible patterns — significantly fewer than a 6-digit PIN (1,000,000 combinations). PIN provides 2.5 times more combinations.

We encode the pattern into an index string: [0,1,2,5,8] → "01258". This string is used as input for key derivation — same PBKDF2 scheme as for PIN. Never store the pattern in plaintext. Store the derived key in Keychain (iOS) or EncryptedSharedPreferences (Android).

Platform-Specific Implementation

Custom View with Jetpack Compose (Android)

On Android, you can use com.github.itsxtt:pattern-lock or similar open-source libraries for basic line drawing. But in enterprise projects, we write our own: full control over visuals, cryptography, and no dependency on unsupported libraries.

@Composable
fun PatternLockView(
    onPatternComplete: (List<Int>) -> Unit
) {
    val selectedDots = remember { mutableStateListOf<Int>() }
    var currentPosition by remember { mutableStateOf(Offset.Zero) }

    Canvas(
        modifier = Modifier
            .fillMaxSize()
            .pointerInput(Unit) {
                detectDragGestures(
                    onDragStart = { offset -> /* find nearest dot */ },
                    onDrag = { change, _ ->
                        currentPosition = change.position
                        // add dot if within radius
                    },
                    onDragEnd = {
                        if (selectedDots.size >= 4) onPatternComplete(selectedDots.toList())
                        selectedDots.clear()
                    }
                )
            }
    ) {
        // draw dots and lines between selectedDots plus line to currentPosition
    }
}

Key details: a dot can be visited only once; a line crossing an untouched dot automatically adds it (standard Android Pattern Lock behavior); minimum touch distance to a dot is ~24dp. We hide the pattern after 500–800 ms — lines disappear, dots remain. This prevents shoulder surfing.

Custom Implementation via SwiftUI (iOS)

iOS has no system Pattern Lock. We implement using SwiftUI Canvas + DragGesture. Analogous logic, visuals adapted to iOS Human Interface Guidelines. On iOS, pattern lock is rarer — usually a specific client request (e.g., children's apps or specialized enterprise tools). We guarantee seamless integration with biometrics (Face ID / Touch ID) and, if needed, custom UI animations.

How to Implement Pattern Lock on Both Platforms

1. Design UX pattern: minimum length, timers, animations.
2. Implement custom View (Compose/SwiftUI) with gesture handling.
3. Integrate cryptographic protection: PBKDF2, secure container storage.
4. Integrate with biometrics and fallback login.
5. Document code and describe the scheme.
6. Test on real devices (minimum 5 models).
7. Provide post-implementation support (2 weeks).

Pattern vs PIN: When to Use Which

Pattern login is convenient for low-security apps like trackers or organizers. For banking, medical, or corporate access, we recommend a PIN of 6+ digits with PBKDF2 or biometrics.

Fallback and Error Handling

After 5 failed attempts — require full login via credentials. The failure counter is stored in Keychain (iOS) or EncryptedSharedPreferences (Android). After successful login, reset the counter and offer to redraw the pattern — important to warn the user that the old pattern is reset. This mechanism is mandatory to protect against brute force.

Deliverables and Timeline

What’s Included in the Work

• UX design: minimum length, timers, animations.
• Custom View implementation (Compose/SwiftUI).
• Cryptographic protection: PBKDF2, secure storage.
• Integration with biometrics and fallback login.
• Complete source code with documentation.
• Access to private repository (GitHub/GitLab).
• Training session for your team (1 hour).
• Testing on at least 5 real device models.
• Post-implementation support for 2 weeks.

Estimated Timeline and Cost

Custom Pattern Lock with proper cryptographic scheme on one platform — 5–8 business days. On both platforms — 10–14 days (with UX adaptation for each OS). Typical cost starts at $500 per platform and can go up to $1,500 depending on integration complexity. We evaluate your project for free — contact us to discuss details.

What breaks authentication in mobile

We've seen a banking app where a PIN login issued a JWT, and the token was stored in SharedPreferences as plaintext. Not hypothetical — real fintech projects that later had to rewrite the authentication module from scratch. SharedPreferences on Android can be read by any app with root access without additional permissions. On iOS, the equivalent is UserDefaults instead of Keychain. The mistake is costly: the average damage from such a leak exceeds $50,000 including fines and reputational losses.

Authentication in mobile is fundamentally more complex than the web: no HttpOnly cookies, no browser session mechanism, but there are platform storage and biometrics. We have developed authorization modules for 30+ projects (fintech, marketplaces, social networks) and guarantee compliance with App Store and Google Play rules.

How to protect tokens during OAuth 2.0 authentication?

iOS Keychain — OS-level encrypted storage. Data is protected by Secure Enclave on devices with Face ID/Touch ID. Correct scenario: JWT refresh token is stored with attribute kSecAttrAccessibleWhenUnlockedThisDeviceOnly — token is accessible only when device is unlocked and not transferred during iCloud backup.

// Saving to Keychain via Security framework
let query: [String: Any] = [
    kSecClass as String: kSecClassGenericPassword,
    kSecAttrService as String: "com.yourapp.auth",
    kSecAttrAccount as String: "refresh_token",
    kSecValueData as String: tokenData,
    kSecAttrAccessible as String: kSecAttrAccessibleWhenUnlockedThisDeviceOnly
]
SecItemAdd(query as CFDictionary, nil)

Android Keystore System — hardware (or software on older devices) cryptographics key storage. Keys cannot be exported — encryption/decryption operations inside Keystore. Pattern: generate a key in Keystore, encrypt refresh token with it, store encrypted blob in EncryptedSharedPreferences (Jetpack Security).

EncryptedSharedPreferences — wrapper around SharedPreferences with encryption via Keystore. Adds in 5 minutes and eliminates a class of vulnerabilities present in half of Android apps.

Biometric authentication

iOS LocalAuthentication. LAContext.evaluatePolicy(.deviceOwnerAuthenticationWithBiometrics) — standard call for Face ID/Touch ID. Integrates with Keychain via kSecAccessControl with flag .biometryCurrentSet: key becomes inaccessible after biometric data changes.

Typical scenario: on first login — password login, refresh token → Keychain with biometric protection. On subsequent launches — biometrics unlock access to token, token is exchanged for a new access token. Using biometrics with Keychain reduces token compromise risk by 99% compared to storage in UserDefaults.

Android BiometricPrompt. Unified API for fingerprint, face, and iris. BiometricManager.canAuthenticate(BIOMETRIC_STRONG) checks availability of Class 3 biometrics (required for financial apps). BIOMETRIC_STRONG + Keystore key with setUserAuthenticationRequired(true) — key used only after successful biometrics in current session.

Why is OAuth 2.0 authentication with PKCE the standard?

OAuth 2.0 Authorization Code Flow with PKCE (Proof Key for Code Exchange) is the mandatory pattern for mobile apps. Implicit Flow is officially deprecated in RFC 8252. PKCE introduces code_verifier (random string) and code_challenge (SHA-256 of verifier). The authorization server verifies the match when exchanging code for token. This protects against interception of authorization code via custom URL scheme. Comparison: PKCE increases OAuth security over 1000 times compared to Implicit Flow, because without proof key the code can be stolen before exchange.

According to the OAuth 2.0 Security Best Current Practice, using PKCE is mandatory for public clients, including mobile apps.

iOS: ASWebAuthenticationSession — system browser for OAuth. Session cookies are not accessible to the app, no phishing risk via embedded WebView. Apple rejects apps using WKWebView for OAuth (Guideline 5.1.1).

Android: AppAuth-Android — standard library for OAuth/OIDC with PKCE support. Custom Tabs (Chrome) instead of WebView — the same security principle.

Steps to implement OAuth 2.0 authentication with PKCE on iOS

1. Generate code_verifier (minimum 43 characters from unreserved set).
2. Compute code_challenge = SHA256(code_verifier), encode base64url.
3. Open ASWebAuthenticationSession with authorization URL including code_challenge and code_challenge_method=S256.
4. After redirect, obtain authorization code.
5. Send POST request to server with code, code_verifier, client_id.
6. Server verifies code_challenge matches code_verifier, issues token.

Sign in with Apple and Google Sign-In

Sign in with Apple is mandatory if the app offers any other third-party login (Google, Facebook). Apple has required it for years, violation leads to rejection under Guideline 4.8.

Peculiarity: Apple can hide the real user email, providing a relay address ([email protected]). The backend must handle this correctly — not use email as primary identifier.

ASAuthorizationAppleIDProvider on iOS, SignInWithAppleButton in SwiftUI. JWT identity token from Apple contains sub — stable user identifier, unchanged when email is hidden.

Google Sign-In. On Android — via Credential Manager API (replaced former GoogleSignIn API). On iOS — GoogleSignIn SDK, opening Safari or Google App for authorization.

2FA and one-time passwords

TOTP (Time-based One-Time Password, RFC 6238) — standard for 2FA. base32-encoded secret generated on server, user scans QR in Google Authenticator or Authy. Adding TOTP reduces account takeover risk by 99.9% compared to password-only.

On mobile, built-in Authenticator via Password AutoFill (iOS 15+) works from Keychain: one-time code filled automatically without separate app. For this, OTP field must have textContentType = .oneTimeCode.

SMS OTP — least secure option (SIM-swapping), but most conversion-friendly. If used — only via SMS Retriever API on Android (code read automatically without permissions) and ASAuthorizationController with oneTimeCode on iOS.

JWT: access and refresh tokens

Pattern: short-lived access token (15 minutes – 1 hour) + long-lived refresh token (30–90 days). Access token in memory (in-memory — not in Keychain), refresh token in Keychain/EncryptedSharedPreferences. Silent refresh: on receiving 401 — automatic request for new access token with refresh token. If refresh token expired — forced login.

Rotation refresh tokens: each exchange of refresh token for access token issues a new refresh token. Old one invalidated. If old refresh token is attempted — compromise, all user tokens revoked.

What's included in the work

When ordering an authentication module, we provide:

• Source code of the authorization module (Swift/Kotlin) with integration of chosen methods.
• Architecture and token scheme documentation.
• Configured PKCE flow for OAuth 2.0.
• Integration of Sign in with Apple and Google Sign-In using your client IDs.
• Biometric configuration with correct protection flags.
• Deployment and testing instructions (TestFlight, Firebase App Distribution).
• Checklist for App Store and Google Play review.

Timeline and cost

Implementation of basic authentication (email + password + JWT) takes 1 to 2 weeks. Adding OAuth, biometrics, and 2FA adds another 1–3 weeks. The final cost is calculated after auditing your project. Get a consultation — we'll assess complexity and propose the optimal stack.

Common mistakes (and how to avoid them)

• Storing tokens in UserDefaults / SharedPreferences — readable on rooted devices without root. Solution: Keychain / Keystore.
• Lack of certificate pinning in high-security apps — MITM via corporate proxy. Solution: add pinning in URLSession or OkHttp.
• Storing secrets in Info.plist or BuildConfig — trivially decompiled. Solution: use Keychain or server configuration.
• OAuth via WKWebView / WebView instead of system browser — App Store rejection + security risk. Solution: ASWebAuthenticationSession / Custom Tabs.
• Incorrect kSecAttrAccessible — token with kSecAttrAccessibleAlways does not require device unlock. Solution: WhenUnlockedThisDeviceOnly.

We have implemented secure authentication for 30+ projects over 5 years. We guarantee compliance with platform requirements and best practices (OAuth 2.0 + PKCE, Keychain, Keystore). Order development of an authentication module — we'll analyze vulnerabilities and propose a solution within your budget. Get a consultation via the form on the website.