Android has built-in security features that significantly reduce the frequency and impact of application security issues. The system is designed so that you can typically build your apps with the default system and file permissions and avoid difficult decisions about security.
The following core security features help you build secure apps:
- The Android application sandbox, which isolates your app data and code execution from other apps.
- An application framework with robust implementations of common security functionality such as cryptography, permissions, and secure interprocess communication (IPC).
- Technologies like address space layout randomization (ASLR),
no-execute (NX), ProPolice, safe_iop,
calloc, and Linux
mmap_min_addrto mitigate risks associated with common memory management errors.
- User-granted permissions to restrict access to system features and user data.
- Application-defined permissions to control application data on a per-app basis.
It's important to be familiar with the Android security best practices on this page. Following these practices as general coding habits help you avoid inadvertently introducing security issues that adversely affect your users.
Authentication is a prerequisite for many key security operations. To control access to protected assets like user data, app functionality, and other resources, you'll need to add authentication to your Android app.
You can improve your user's authentication experience by integrating your app with Credential Manager. Credential Manager is an Android Jetpack library that unifies API support for most major authentication methods, including passkeys, passwords, and federated sign-in solutions such as Sign-in with Google.
To further enhance security for your app, consider adding biometric authentication methods such as fingerprint scans or facial recognition. Good candidates for adding biometric authentication might include apps for financial, healthcare, or identity management.
Android's autofill framework can ease the sign-up and sign-in process, reducing error rates and user friction. Autofill integrates with password managers, allowing users to select complex, randomized passwords that can be stored and retrieved easily and securely.
The most common security concern for an application on Android is whether the data that you save on the device is accessible to other apps. There are three fundamental ways to save data on the device:
- Internal storage
- External storage
- Content providers
The following sections describe the security issues associated with each approach.
By default, files that you create on internal storage are accessible only to your app. Android implements this protection, and it's sufficient for most applications.
Avoid the deprecated
MODE_WORLD_READABLE modes for IPC files. They don't provide the ability
to limit data access to particular applications, and they don't provide any
control of data format. If you want to share your data with other app processes,
consider using a content provider instead, which offers read and write
permissions to other apps and can make dynamic permission grants on a
Files created on external storage, such as SD cards, are globally readable and writable. Because external storage can be removed by the user and also modified by any application, only store non-sensitive information using external storage.
Perform input validation when handling data from external storage as you would with data from any untrusted source. Don't store executables or class files on external storage prior to dynamic loading. If your app does retrieve executable files from external storage, make sure the files are signed and cryptographically verified prior to dynamic loading.
Content providers offer a structured storage mechanism that can be limited
to your own application or exported to allow access by other applications. If
you don't intend to provide other applications with access to your
ContentProvider, mark it as
android:exported=false in the
application manifest. Otherwise, set the
android:exported attribute to
to let other apps access the stored data.
When creating a
ContentProvider that is exported for use by other
applications, you can specify a single permission for reading and writing,
or you can specify distinct permissions for reading and writing. Limit your
permissions to those required to accomplish the task at hand. Keep in mind that
it's usually easier to add permissions later to expose new functionality than it
is to take them away and impact existing users.
If you are using a content provider for sharing data between only your own apps,
we recommend using the
android:protectionLevel attribute set to
signature protection. Signature permissions don't require user
confirmation, so they provide a better user experience and more controlled
access to the content provider data when the apps accessing the data are
signed with the same key.
Content providers can also provide more granular access by declaring the
android:grantUriPermissions attribute and using the
FLAG_GRANT_WRITE_URI_PERMISSION flags in the
Intent object that
activates the component. The scope of these permissions can be further limited
When accessing a content provider, use parameterized query methods such as
delete() to avoid potential SQL
injection from untrusted sources. Note that using parameterized methods is not
sufficient if the
selection argument is built by concatenating user data prior
to submitting it to the method.
Don't have a false sense of security about the write permission. The write
permission allows SQL statements that make it possible for some data to be
confirmed using creative
WHERE clauses and parsing the results. For example,
an attacker might probe for the presence of a specific phone number in a call
log by modifying a row only if that phone number already exists. If the content
provider data has predictable structure, the write permission might be
equivalent to providing both reading and writing.
Because Android sandboxes applications from each other, applications must explicitly share resources and data. They do this by declaring the permissions they need for additional capabilities not provided by the basic sandbox, including access to device features such as the camera.
Minimize the number of permissions that your app requests. Restricting access to sensitive permissions reduces the risk of inadvertently misusing those permissions, improves user adoption, and makes your app less vulnerable for attackers. Generally, if a permission isn't required for your app to function, don't request it. See the guide to evaluating whether your app needs to declare permissions.
If possible, design your application in a way that doesn't require any permissions. For example, rather than requesting access to device information to create a unique identifier, create a UUID for your application. (Learn more in the section about user data). Or, rather than using external storage (which requires permission), store data on internal storage.
In addition to requesting permissions, your application can use the
<permission> element to protect IPC that is security sensitive and is
exposed to other applications, such as a
ContentProvider. In general, we
recommend using access controls other than user-confirmed permissions where
possible, because permissions can be confusing for users. For example, consider
using the signature protection level on permissions for IPC communication
between applications provided by a single developer.
Don't leak permission-protected data. This occurs when your app exposes data over IPC that is available only because your app has permission to access that data. The clients of your app's IPC interface might not have that same data-access permission. More details on the frequency and potential effects of this issue appear in the research paper Permission Re-Delegation: Attacks and Defenses , published at USENIX.
Define the smallest set of permissions that satisfy your security requirements. Creating a new permission is relatively uncommon for most applications, because the system-defined permissions cover many situations. Where appropriate, perform access checks using existing permissions.
If you need a new permission, consider whether you can accomplish your task with a signature protection level. Signature permissions are transparent to the user and allow access only by applications signed by the same developer as the application performing the permission check.
If creating a new permission is still required, declare it in the app manifest
<permission> element. Apps using the new permission can
reference it by adding a
<uses-permission> element in their manifest
files. You can also add permissions dynamically by using the
If you create a permission with the dangerous protection level, there are a number of complexities that you need to consider:
- The permission must have a string that concisely expresses to the user the security decision they are required to make.
- The permission string must be localized to many different languages.
- Users might choose not to install an application because a permission is confusing or perceived as risky.
- Applications might request the permission when the creator of the permission hasn't been installed.
Each of these poses a significant nontechnical challenge for you as the developer while also confusing your users, which is why we discourage the use of the dangerous permission level.
Network transactions are inherently risky for security, because they involve transmitting data that is potentially private to the user. People are increasingly aware of the privacy concerns of a mobile device, especially when the device performs network transactions, so it's very important that your app implement all best practices toward keeping the user's data secure at all times.
Networking on Android is not significantly different from other Linux
environments. The key consideration is making sure that appropriate protocols
are used for sensitive data, such as
HttpsURLConnection for secure web
traffic. Use HTTPS over HTTP anywhere that HTTPS is supported on the server,
because mobile devices frequently connect on networks that aren't secured, such
as public Wi-Fi hotspots.
Authenticated, encrypted socket-level communication can be easily implemented
SSLSocket class. Given the frequency with which Android
devices connect to unsecured wireless networks using Wi-Fi, the use of secure
networking is strongly encouraged for all applications that communicate over the
Some applications use localhost network ports for handling
sensitive IPC. Don't use this approach, because these interfaces are accessible
by other applications on the device. Instead, use an Android IPC mechanism where
authentication is possible, such as with a
Service. Binding to the
non-specific IP address
INADDR_ANY is worse than using loopback, because it
allows your application to receive requests from any IP address.
Make sure that you don't trust data downloaded from HTTP or other insecure
protocols. This includes validation of input in
WebView and any
responses to intents issued against HTTP.
The Short Message Service (SMS) protocol was primarily designed for user-to-user communication and isn't well suited for apps that want to transfer data. Due to the limitations of SMS, we recommend using Firebase Cloud Messaging (FCM) and IP networking for sending data messages from a web server to your app on a user device.
Be aware that SMS is neither encrypted nor strongly authenticated on either the
network or the device. In particular, any SMS receiver should expect that a
malicious user might have sent the SMS to your application. Don't rely on
unauthenticated SMS data to perform sensitive commands. Also, be aware that SMS
can be subject to spoofing and/or interception on the network. On the
Android-powered device itself, SMS messages are transmitted as broadcast
intents, so they can be read or captured by other applications that have the
Insufficient input validation is one of the most common security problems affecting applications, regardless of what platform they run on. Android has platform-level countermeasures that reduce the exposure of applications to input validation issues, and we recommend that you use those features where possible. Also, we recommend using type-safe languages to reduce the likelihood of input validation issues.
If you are using native code, any data read from files, received over the network, or received from an IPC has the potential to introduce a security issue. The most common problems are buffer overflows, use after free, and off-by-one errors. Android provides a number of technologies, like ASLR and Data Execution Prevention (DEP), that reduce the exploitability of these errors, but they don't solve the underlying problem. You can prevent these vulnerabilities by carefully handling pointers and managing buffers.
If you are using data within queries that are submitted to an SQL database or a content provider, SQL injection can be an issue. The best defense is to use parameterized queries, as discussed in the section about content providers. Limiting permissions to read-only or write-only can also reduce the potential for harm related to SQL injection.
If you can't use the security features discussed in this section, make sure to use well-structured data formats and verify that the data conforms to the expected format. While blocking specific characters or performing character replacement can be an effective strategy, these techniques are error prone in practice, and we recommend avoiding them when possible.
The best approach for user data security is to minimize the use of APIs that access sensitive or personal information. If you have access to user data, avoid storing or transmitting it if you can. Consider whether your application logic can be implemented using a hash or non-reversible form of the data. For example, your app might use the hash of an email address as a primary key to avoid transmitting or storing the email address. This reduces the chances of inadvertently exposing data, and it also reduces the chance of attackers attempting to exploit your app.
Also, consider whether your application could inadvertently expose personal information to other parties, such as third-party components for advertising or third-party services used by your application. If you don't know why a component or service requires personal information, don't provide it. In general, reducing the access to personal information by your application reduces the potential for problems in this area.
If your app requires access to sensitive data, evaluate whether you need to transmit it to a server or if you can run the operation on the client. Consider running any code using sensitive data on the client to avoid transmitting user data. Also, make sure that you don't inadvertently expose user data to other applications on the device through overly permissive IPC, world-writable files, or network sockets. Overly permissive IPC is a special case of leaking permission-protected data, discussed in the Permission requests section.
If a Globally Unique Identifier (GUID) is required, create a large, unique number and store it. Don't use phone identifiers such as the phone number or IMEI, which might be associated with personal information. This topic is discussed in more detail in the page about best practices for unique identifiers.
Be careful when writing to on-device logs. On Android, logs are a shared
resource and are available to an application with the
permission. Even though the phone log data is temporary and erased on reboot,
inappropriate logging of user information could inadvertently leak user data to
other applications. In addition ton't logging PII, limit log usage in
production apps. To easily implement this, use debug flags and custom
classes with easily configurable logging levels.
WebView consumes web content that can include HTML and
number of mechanisms to reduce the scope of these potential issues by limiting
the capability of
WebView to the minimum functionality required by your
repurpose sample code that uses it in a production application, remove that
method call if it's not required. By default,
WebView doesn't execute
applications. If you use it, expose
from which all input is trustworthy. If untrusted input is allowed, untrusted
we recommend exposing
contained within your application APK.
If your application accesses sensitive data with a
WebView, consider using the
clearCache() method to delete any files stored locally. You can also use
server-side headers, such as
no-store, to indicate that an application should
not cache particular content.
Devices running platforms older than Android 4.4 (API level 19) use a version of
webkit that has a number of security issues. As a workaround, if your
app is running on these devices, it must confirm that
WebView objects display
only trusted content. To make sure your app isn't exposed to potential
vulnerabilities in SSL, use the updatable security
Provider object as
described in Update your security provider to protect against SSL
exploits. If your application must render content from the open web,
consider providing your own renderer so you can keep it up to date with the
latest security patches.
Credential requests are a vector for attack. Here are some tips to help you make credential requests in your Android apps more secure.
Minimize credential exposure
- Avoid unnecessary credential requests. To make phishing attacks more conspicuous and less likely to be successful, minimize the frequency of asking for user credentials. Instead, use an authorization token and refresh it. Request only the minimum amount of credential information necessary for authentication and authorization.
- Store credentials securely. Use Credential Manager to enable passwordless authentication using passkeys or to implement federated sign-in using schemes such as Sign in with Google. If you must use traditional password authentication, don't store user IDs and passwords on the device. Instead, perform initial authentication using the username and password supplied by the user, and then use a short-lived, service-specific authorization token.
- Limit the scope of permissions. Don't request broad permissions for a task that only requires a more narrow scope.
- Limit access tokens. Use short-lived tokens operations and API calls.
- Limit authentication rates. Rapid, successive authentication or authorization requests can be a sign of a brute-force attack. Limit these rates to a reasonable frequency while still allowing for a functional and user-friendly app experience.
Use secure authentication
- Implement passkeys. Enable passkeys as a more secure and user-friendly upgrade to passwords.
- Add biometrics. Offer the ability to use biometric authentication such as fingerprint or facial recognition for added security.
- Use federated identity providers. Credential Manager supports federated authentication providers such as Sign in with Google.
- Encrypt communication Use HTTPS and similar technologies to ensure the data your app sends over a network is protected.
Practice secure account management
- Connect to services that are accessible to multiple applications using
AccountManager. Use the
AccountManagerclass to invoke a cloud-based service, and don't store passwords on the device.
- After using
AccountManagerto retrieve an
CREATORbefore passing in any credentials so that you don't inadvertently pass credentials to the wrong application.
- If credentials are used only by applications that you create, you can verify
the application that accesses the
checkSignatures. Alternatively, if only one application uses the credential, you might use a
- Keep your code up-to-date. Be sure to update your source code, including any third-party libraries and dependencies, to guard against the latest vulnerabilities.
- Monitor suspicious activity. Look for potential misuse, such as patterns of authorization misuse.
- Audit your code. Perform regular security checks against your codebase to look for potential credential request issues.
API key management
API keys are a critical component of many Android apps, enabling them to access external services and perform essential functions such as connecting to mapping services, authentication, and weather services. However, exposing these sensitive keys can have severe consequences, including data breaches, unauthorized access, and financial losses. To prevent such scenarios, developers should implement secure strategies for handling API keys throughout the development process.
To protect services from misuse, API keys must be carefully protected. To secure a connection between the app and a service that uses an API key, you need to secure the access to the API. When your app is compiled, and your app's source code includes API keys, it's possible for an attacker to decompile the app and find these resources.
This section is intended for two groups of Android developers: those who work with infrastructure teams on their continuous delivery pipeline, and those who deploy standalone apps in the Play store. This section outlines best practices for how to handle API keys, so your app can communicate with services securely.
Generation and storage
Developers should treat API key storage as a critical component of data protection and user privacy using a defense-in-depth approach.
Strong key storage
The following example uses the Jetpack security-crypto library to create encrypted shared preferences.
val masterKey = MasterKey.Builder(context)
val encryptedSharedPreferences = EncryptedSharedPreferences.create(
// use the shared preferences and editor as you normally would
MasterKey masterKey = new MasterKey.Builder(context)
SharedPreferences sharedPreferences = EncryptedSharedPreferences.create(
// use the shared preferences and editor as you normally would
SharedPreferences.Editor editor = sharedPreferences.edit();
Source control exclusion
Never commit API keys to your source code repository. Adding API keys to source code risks exposure of keys to public repositories, shared code examples, and accidentally-shared files. Instead, use Gradle plugins such as the secrets-gradle-plugin to work with API keys in your project.
If possible, use separate API keys development, testing, and production environments. Use environment-specific keys to isolate each environment, reducing the risk of exposing production data and allowing you to disable compromised keys without affecting your production environment.
Usage and access control
Secure API key practices are essential for protecting your API and your users. Here's how to prepare your keys for optimal security:
- Generate unique keys for each app: Use separate API keys for each app to help identify and isolate compromised access.
- Implement IP restrictions: If possible, limit API key usage to specific IP addresses or ranges.
- Limit mobile app key usage: Limit API key usage to specific mobile apps by bundling them with the key or by using app certificates.
- Log and monitor for suspicious activity: Implement API usage logging and monitoring mechanisms to detect suspicious activity and prevent potential abuse.
Note: Your service should provide features for restricting keys to a particular package or platform. For example, the Google Maps API limits key access by package name and signing key.
OAuth 2.0 provides a framework for authorizing access to resources. It defines standards for how clients and servers should interact, and it allows for secure authorization. You can use OAuth 2.0 to restrict API key usage to specific clients, and define the access scope so that each API key only has the minimum level of access required for their intended purpose.
Key rotation and expiration
To reduce the risk of unauthorized access through undiscovered API vulnerabilities, it is important to rotate API keys regularly. The ISO 27001 standard defines a compliance framework for how often to perform key rotation. For most cases, a key rotation period between 90 days to 6 months should be adequate. Implementing a robust key management system can help you streamline these processes, improving the efficiency of your key rotation and expiration needs.
General best practices
- Use SSL/HTTPS: Always use HTTPS communication to encrypt your API requests.
- Certificate pinning: For an extra layer of security, you can consider implementing certificate pinning to check which certificates are considered valid.
- Validate and sanitize user input: Validate and sanitize user input to prevent injection attacks that could expose API keys.
- Follow security best practices: Implement general security best practices in your development process, including secure coding techniques, code reviews, and vulnerability scanning.
- Stay informed: Stay updated on the latest security threats and best practices for API key management.
- SDKs up-to-date: Make sure your SDKs and libraries are updated to the latest version.
In addition to providing data isolation, supporting full-filesystem encryption, and providing secure communications channels, Android provides a wide array of algorithms for protecting data using cryptography.
Know which Java Cryptography Architecture (JCA) security providers your software uses. Try to use the highest level of the pre-existing framework implementation that can support your use case. If applicable, use the Google-provided providers in the Google-specified order.
If you need to securely retrieve a file from a known network location, a simple
HTTPS URI might be adequate and requires no knowledge of cryptography. If you
need a secure tunnel, consider using
SSLSocket rather than writing your own protocol. If you use
be aware that it doesn't perform hostname verification. See Warnings about
If you find that you need to implement your own protocol, don't implement your
own cryptographic algorithms. Use existing cryptographic algorithms, such as the
implementations of AES and RSA provided in the
Additionally, follow these best practices:
- Use 256-bit AES for commercial purposes. (If unavailable, use 128-bit AES.)
- Use either 224- or 256-bit public key sizes for elliptic curve (EC) cryptography.
- Know when to use CBC, CTR, or GCM block modes.
- Avoid IV/counter reuse in CTR mode. Ensure that they're cryptographically random.
- When using encryption, implement integrity using the CBC or CTR mode with one
of the following functions:
- GCM mode
Use a secure random number generator,
SecureRandom, to initialize any
cryptographic keys generated by
KeyGenerator. Use of a key that is not
generated with a secure random number generator significantly weakens the
strength of the algorithm and may allow offline attacks.
If you need to store a key for repeated use, use a mechanism, such as
KeyStore, that provides long term storage and retrieval of cryptographic
Some apps attempt to implement IPC using traditional Linux techniques such as
network sockets and shared files. However, we recommend instead that you use
Android system functionality for IPC such as
Messenger with a
Android IPC mechanisms let you verify the identity of the application connecting
to your IPC and set security policy for each IPC mechanism.
Many of the security elements are shared across IPC mechanisms. If your IPC
mechanism isn't intended for use by other applications, set the
android:exported attribute to
false in the component's manifest element,
such as for the
<service> element. This is useful for applications that
consist of multiple processes within the same UID or if you decide late in
development that you don't actually want to expose functionality as IPC, but you
don't want to rewrite the code.
If your IPC is accessible to other applications, you can apply a security policy
by using the
<permission> element. If the IPC is between apps that are
your own and are signed with the same key, use a
signature-level permission in
For activities and broadcast receivers, intents are the preferred mechanism for
asynchronous IPC on Android. Depending on your application requirements, you
sendOrderedBroadcast, or an explicit
intent to a specific application component. For security purposes, explicit
intents are preferred.
Caution: If you use an intent to bind to a
**Service**, use an
explicit intent to keep your app secure. Using an implicit intent to start
a service is a security hazard, because you can't be certain what service will
respond to the intent and the user can't see which service starts. Beginning
with Android 5.0 (API level 21), the system throws an exception if you call
**bindService()** with an implicit intent.
Note that ordered broadcasts can be consumed by a recipient, so they might not be delivered to all applications. If you are sending an intent that must be delivered to a specific receiver, you must use an explicit intent that declares the receiver by name.
Senders of an intent can verify that the recipient has permission by specifying a non-null permission with the method call. Only applications with that permission receive the intent. If data within a broadcast intent might be sensitive, consider applying a permission to make sure that malicious applications can't register to receive those messages without appropriate permissions. In those circumstances, you might also consider invoking the receiver directly, rather than raising a broadcast.
Note: Intent filters aren't security features. Components can be invoked with explicit intents and might not have data that would conform to the intent filter. To confirm that it is properly formatted for the invoked receiver, service, or activity, perform input validation within your intent receiver.
By default, services aren't exported and can't be invoked by any other
application. However, if you add any intent filters to the service declaration,
it is exported by default. It's best if you explicitly declare the
android:exported attribute to be sure it behaves the way you intend it
to. Services can also be protected using the
attribute. By doing so, other applications need to declare a corresponding
<uses-permission> element in their own manifest to be able to start,
stop, or bind to the service.
Note: If your app targets Android 5.0 (API level 21) or higher, use the
**JobScheduler** to execute background services.
A service can protect individual IPC calls that are made into it with
permissions. This is done by calling
executing the implementation of the call. We recommend using the declarative
permissions in the manifest, since those are less prone to oversight.
Caution: Don't confuse client and server permissions; ensure that the called app has appropriate permissions and verify that you grant the same permissions to the calling app.
Binder and Messenger interfaces
We recommend that you design your app interfaces in a way that doesn't require
interface-specific permission checks.
Messenger objects aren't
declared within the application manifest, and therefore you can't apply
declarative permissions directly to them. They generally inherit permissions
declared in the application manifest for the
within which they are implemented. If you are creating an interface that
requires authentication and/or access controls, you must explicitly add those
controls as code in the
If you are providing an interface that does require access controls, use
checkCallingPermission() to verify whether the caller has a required
permission. This is especially important before accessing a service on behalf of
the caller, as the identity of your application is passed to other interfaces.
If you are invoking an interface provided by a
bindService() invocation can fail if you don't have permission to access
the given service. If you need to allow an external process to interact with
your app but it doesn't have the necessary permissions to do so, you can use the
clearCallingIdentity() method. This method performs the call to your
app's interface as though your app were making the call itself, rather than the
external caller. You can restore the caller permissions later with the
For more information about performing IPC with a service, see Bound Services.
By default, receivers are exported and can be invoked by any other application.
BroadcastReceiver is intended for use by other applications, you might
want to apply security permissions to receivers using the
element within the application manifest. This prevents applications without
appropriate permissions from sending an intent to the
Security with dynamically loaded code
We strongly discourage loading code from outside of your application APK. Doing so significantly increases the likelihood of application compromise due to code injection or code tampering. It also adds complexity around version management and application testing—and it can make it impossible to verify the behavior of an application, so it might be prohibited in some environments.
If your application does dynamically load code, the most important thing to keep in mind is that the dynamically loaded code runs with the same security permissions as the application APK. The user makes a decision to install your application based on your identity, and the user expects that you provide any code run within the application, including code that is dynamically loaded.
Many applications attempt to load code from insecure locations, such as
downloaded from the network over unencrypted protocols or from world-writable
locations such as external storage. These locations could let someone on the
network modify the content in transit or another application on a user's device
to modify the content on the device. On the other hand, modules included
directly within your APK can't be modified by other applications. This is true
whether the code is a native library or a class being loaded using
Security in a virtual machine
Dalvik is Android's runtime virtual machine (VM). Dalvik was built specifically for Android, but many of the concerns regarding secure code in other virtual machines also apply to Android. In general, you don't need to concern yourself with security issues relating to the virtual machine. Your application runs in a secure sandbox environment, so other processes on the system can't access your code or private data.
If you're interested in learning more about virtual machine security, familiarize yourself with some existing literature on the subject. Two of the more popular resources are:
This document focuses on areas that are Android specific or different from other VM environments. For developers experienced with VM programming in other environments, there are two broad issues that might be different about writing apps for Android:
- Some virtual machines, such as the JVM or .NET runtime, act as a security boundary, isolating code from the underlying operating system capabilities. On Android, the Dalvik VM is not a security boundary—the application sandbox is implemented at the OS level, so Dalvik can interoperate with native code in the same application without any security constraints.
- Given the limited storage on mobile devices, it's common for developers to want to build modular applications and use dynamic class loading. When doing this, consider both the source where you retrieve your application logic and where you store it locally. Don't use dynamic class loading from sources that aren't verified, such as unsecured network sources or external storage, because that code might be modified to include malicious behavior.
Security in native code
In general, we recommend using the Android SDK for application development, rather than using native code with the Android NDK. Applications built with native code are more complex, less portable, and more likely to include common memory-corruption errors such as buffer overflows.
Android is built using the Linux kernel, and being familiar with Linux development security best practices is especially useful if you are using native code. Linux security practices are beyond the scope of this document, but one of the most popular resources is Secure Programming HOWTO - Creating Secure Software.
An important difference between Android and most Linux environments is the application sandbox. On Android, all applications run in the application sandbox, including those written with native code. A good way to think about it for developers familiar with Linux is to know that every application is given a unique User Identifier (UID) with very limited permissions. This is discussed in more detail in the Android Security Overview, and you should be familiar with application permissions even if you are using native code.