security-talent/bugskill-ai

mobile-network-security

Detects insecure network communication in mobile apps (Android/iOS).

ソースを見る

使い方

  1. インストールコマンドをコピーし、自分の環境へ導入する前にソースを確認します。
  2. 必要なツール、認証情報、対応環境を原文で確認します。
  3. 入力と成果物を指定し、出力を資料と照合します。
npx skills add https://github.com/security-talent/bugskill-ai --skill mobile-network-security

開始前の確認

  • 導入にはモデル契約、API枠、ファイルへのアクセス許可は含まれません。
  • スター数や導入数は活動の指標であり、独立した品質・安全性試験ではありません。
リポジトリの原文

見出し、例、コード、表、リンク、参照画像を含む原文を表示しています。

Mobile Network Security

What Is Broken and Why

Mobile apps fail network security when they allow cleartext HTTP traffic, disable TLS certificate validation, or implement certificate pinning incorrectly. Custom X509TrustManager implementations that accept all certificates (empty checkServerTrusted) are a common developer shortcut that makes the entire TLS layer useless. ATS exceptions in iOS Info.plist or Android Network Security Configuration that allow arbitrary cleartext expose all traffic to MITM. Apps that call onReceivedSslError().proceed() in WebViewClient bypass all certificate errors. Certificate pinning without key backup pins causes production outages, so developers remove pinning — leaving no protection.

Key Signals

  • Android: android:networkSecurityConfig pointing to XML with <domain-config cleartextTrafficPermitted="true">
  • Android: android:usesCleartextTraffic="true" in manifest
  • iOS: NSAllowsArbitraryLoads: true in Info.plist ATS section
  • Custom X509TrustManager with empty checkServerTrusted() method body
  • HostnameVerifier returning true for all hosts: ALLOW_ALL_HOSTNAME_VERIFIER
  • SSLContext.init(null, arrayOf(trustAllManager), null)
  • WebViewClient onReceivedSslError calling handler.proceed()
  • TLS 1.0/1.1 explicitly enabled via SSLParameters.setProtocols()
  • No pin-set in Network Security Configuration for sensitive domains
  • iOS NSURLSessionDelegate returning no error for invalid certificates
  • URLSession.shared with no custom delegate (no pinning) for high-value endpoints

Methodology

Setup MITM proxy:

  1. Install Burp/mitmproxy CA cert on device (Android: Settings > Security; iOS: Settings > General > VPN & Device Management)
  2. Configure device proxy to point at Burp listener
  3. Launch app — observe if traffic appears in proxy (cleartext) or throws certificate errors (pinning)

Android static analysis:

  1. apktool d app.apk — check AndroidManifest.xml for usesCleartextTraffic, networkSecurityConfig
  2. Review res/xml/network_security_config.xml for cleartext rules and pin-set presence
  3. Search decompiled source for TrustManager, HostnameVerifier, ALLOW_ALL, onReceivedSslError
  4. Search for SSLContext.init, HttpsURLConnection.setDefaultHostnameVerifier
  5. Check OkHttp client config: OkHttpClient.Builder() for custom sslSocketFactory

iOS static analysis:

  1. Extract IPA — inspect Info.plist for NSAppTransportSecurity exceptions
  2. Search source for URLSession, NSURLConnection, custom URLSessionDelegate methods
  3. Check didReceiveChallenge delegate for completionHandler(.useCredential, ...)
  4. Look for TrustKit, Alamofire, or custom pinning implementation

Dynamic analysis:

  1. With Burp proxy active — if app connects normally: no pinning or pinning bypass available
  2. Attempt SSL kill switch: objection ios sslpinning disable or Android android sslpinning disable
  3. Use Frida script to hook TrustManagerImpl.checkServerTrusted or SecTrustEvaluate

Payloads & Tools

bash
# objection — disable SSL pinning (Android/iOS)
objection --gadget TARGET run android sslpinning disable
objection --gadget TARGET run ios sslpinning disable

# Frida — Android: bypass TrustManager
Java.perform(function() {
  var TrustManager = Java.use("javax.net.ssl.X509TrustManager");
  var SSLContext = Java.use("javax.net.ssl.SSLContext");
  var TM = Java.registerClass({
    name: "FakeTrustManager", implements: [TrustManager],
    methods: { checkClientTrusted: function(){}, checkServerTrusted: function(){},
               getAcceptedIssuers: function(){ return []; } }
  });
  SSLContext.init.overload("[Ljavax.net.ssl.KeyManager;","[Ljavax.net.ssl.TrustManager;","java.security.SecureRandom")
    .implementation = function(km, tm, sr) { this.init(km, [TM.$new()], sr); };
});

# iOS — SSL kill switch (jailbroken device)
# Install SSL Kill Switch 3 via Cydia/Sileo
# OR use Frida script ssl-kill-switch2.js

# Check ATS config in IPA
unzip app.ipa; grep -A20 "NSAppTransportSecurity" Payload/App.app/Info.plist

# Check Android Network Security Config
apktool d app.apk && cat app/res/xml/network_security_config.xml

Bypass Techniques

  • objection sslpinning disable — hooks common pinning libraries (OkHttp, TrustKit, Alamofire) at runtime
  • SSL Kill Switch 3 — jailbroken iOS; patches SecTrustEvaluate at OS level
  • Frida TrustManager replacement — replaces the app's trust manager with one that accepts all certs
  • MagiskTrustUserCerts — on rooted Android, installs CA cert as system cert (bypasses Android 14+ restrictions)
  • apk-mitm — patches APK to disable pinning statically without needing runtime instrumentation
  • Network Security Config override — repack APK with cleartextTrafficPermitted="true" and custom trust anchors

Exploitation Scenarios

Scenario 1 — Empty TrustManager MITM Setup: App uses SSLContext.init(null, arrayOf(TrustAllManager()), null) to avoid pinning errors in dev, shipped to production. → Trigger: Attacker on same Wi-Fi runs mitmproxy. → Impact: All HTTPS traffic decrypted — credentials, session tokens, PII visible.

Scenario 2 — ATS Exception Cleartext Setup: iOS app sets NSAllowsArbitraryLoads: true for legacy API compatibility. → Trigger: Network interception on hotel Wi-Fi. → Impact: Plaintext auth tokens and API responses captured.

Scenario 3 — WebView onReceivedSslError Bypass Setup: WebViewClient overrides onReceivedSslError and calls handler.proceed(). → Trigger: MITM proxy presents a self-signed cert to the WebView. → Impact: Victim navigates authenticated WebView session through attacker's proxy.

False Positives

  • cleartextTrafficPermitted="true" only for non-sensitive domains (analytics, CDN assets) with sensitive traffic separately pinned
  • Custom URLSessionDelegate that validates the cert chain manually and only accepts the prod CA
  • Debug-only bypass code that is stripped in release builds (BuildConfig.DEBUG guard)
  • Certificate pinning disabled for localhost (test environment) — confirm production build behavior

Fix Patterns

xml
<!-- Android Network Security Config — correct -->
<network-security-config>
  <domain-config>
    <domain includeSubdomains="true">api.TARGET</domain>
    <pin-set expiration="2026-01-01">
      <pin digest="SHA-256">SPKI_HASH_HERE</pin>
      <pin digest="SHA-256">BACKUP_SPKI_HASH</pin>  <!-- Always include backup pin -->
    </pin-set>
  </domain-config>
</network-security-config>
swift
// iOS — URLSession pinning via TrustKit or manual
func urlSession(_ session: URLSession, didReceive challenge: URLAuthenticationChallenge,
    completionHandler: @escaping (URLSession.AuthChallengeDisposition, URLCredential?) -> Void) {
    guard let serverTrust = challenge.protectionSpace.serverTrust,
          validateCert(serverTrust) else {   // compare SPKI hash
        completionHandler(.cancelAuthenticationChallenge, nil); return
    }
    completionHandler(.useCredential, URLCredential(trust: serverTrust))
}

Related Skills

[[cors-misconfig]] on mobile backend APIs mirrors the same trust boundary issue as missing certificate pinning — both allow a network-positioned attacker to intercept or manipulate authenticated traffic. An empty TrustManager is functionally equivalent to [[ssrf]] from the attacker's perspective: the server (or in this case the app) makes authenticated requests to an unverified destination. [[mobile-insecure-storage]] is the fallback attack when network interception fails — if TLS is properly pinned, credentials may still be extractable from local storage.

同じリポジトリから

関連する Skills

すべての Skills
security-talent
コミュニティ

401-403-bypass-techniques

401/403 bypass playbook. Use when encountering access-denied responses on admin panels, API endpoints, or restricted paths. Covers path manipulation, HTTP method tampering, header injection, protocol downgrade, and automated bypass tools.

1時間の導入数
1
GitHub Stars
15
更新日
9月26日
security-talent
コミュニティ

active-directory-acl-abuse

Active Directory ACL abuse playbook. Use when exploiting misconfigured AD permissions including GenericAll, WriteDACL, DCSync rights, shadow credentials, LAPS reading, GPO abuse, and BloodHound-guided attack paths.

1時間の導入数
1
GitHub Stars
15
更新日
9月26日
security-talent
コミュニティ

active-directory-certificate-services

AD Certificate Services attack playbook. Use when targeting misconfigured AD CS for privilege escalation via ESC1-ESC13 template abuse, NTLM relay to enrollment, CA officer abuse, and certificate-based persistence.

1時間の導入数
1
GitHub Stars
15
更新日
9月26日
security-talent
コミュニティ

active-directory-kerberos-attacks

Kerberos attack playbook for Active Directory. Use when targeting AD authentication via AS-REP roasting, Kerberoasting, golden/silver/diamond tickets, delegation abuse, or pass-the-ticket attacks.

1時間の導入数
1
GitHub Stars
15
更新日
9月26日