The SWAT Checklist provides an easy to reference set of best practices that raise awareness and help development teams create more secure applications. It's a first step toward building a base of security knowledge around web application security. Use this checklist to identify the minimum standard that is required to neutralize vulnerabilities in your critical applications.

Error Handling and Logging

Best PracticeDescriptionCWE ID
Display Generic Error MessagesError messages should not reveal details about the internal state of the application. For example, file system path and stack information should not be exposed to the user through error messages.cwe-209
No Unhandled ExceptionsGiven the languages and frameworks in use for web application development, never allow an unhandled exception to occur. Error handlers should be configured to handle unexpected errors and gracefully return controlled output to the user.cwe-391
Suppress Framework Generated ErrorsYour development framework or platform may generate default error messages. These should be suppressed or replaced with customized error messages as framework generated messages may reveal sensitive information to the user.cwe-209
Log All Authentication ActivitiesAny authentication activities, whether successful or not, should be logged.cwe-778
Log All Privilege ChangesAny activities or occasions where the user's privilege level changes should be logged.cwe-778
Log Administrative ActivitiesAny administrative activities on the application or any of its components should be logged.cwe-778
Log Access to Sensitive DataAny access to sensitive data should be logged. This is particularly important for corporations that have to meet regulatory requirements like HIPAA, PCI, or SOX.cwe-778
Do Not Log Inappropriate DataWhile logging errors and auditing access is important, sensitive data should never be logged in an unencrypted form. For example, under HIPAA and PCI, it would be a violation to log sensitive data into the log itself unless the log is encrypted on the disk. Additionally, it can create a serious exposure point should the web application itself become compromised.cwe-532
Store Logs SecurelyLogs should be stored and maintained appropriately to avoid information loss or tampering by intruder. Log retention should also follow the retention policy set forth by the organization to meet regulatory requirements and provide enough information for forensic and incident response activities.cwe-533

Data Protection

Best PracticeDescriptionCWE ID
Use HTTPS EverywhereIdeally, HTTPS should be used for your entire application. If you have to limit where it's used, then HTTPS must be applied to any authentication pages as well as to all pages after the user is authenticated. If sensitive information (e.g. personal information) can be submitted before authentication, those features must also be sent over.
Example: Firesheep
Disable HTTP Access for All Protected ResourcesFor all pages requiring protection by HTTPS, the same URL should not be accessible via the insecure HTTP channel.cwe-319
Use the Strict-Transport-Security HeaderThe Strict-Transport-Security header ensures that the browser does not talk to the server over HTTP. This helps reduce the risk of HTTP downgrade attacks as implemented by the sslsniff tool.
Store User Passwords Using a Strong, Iterative, Salted HashUser passwords must be stored using secure hashing techniques with strong algorithms like PBKDF2, bcrypt, or SHA-512. Simply hashing the password a single time does not sufficiently protect the password. Use adaptive hashing (a work factor), combined with a randomly generated salt for each user to make the hash strong.
Example: LinkedIn password leak
Securely Exchange Encryption KeysIf encryption keys are exchanged or pre-set in your application then any key establishment or exchange must be performed over a secure channel
Set Up Secure Key Management ProcessesWhen keys are stored in your system they must be properly secured and only accessible to the appropriate staff on a need to know basis.
Example: AWS Key Management Service (KMS), Azure Key Vault, AWS CloudHSM
Weak TLS Configuration on ServersWeak ciphers must be disabled on all servers. For example, SSL v2, SSL v3, and TLS protocols prior to 1.2 have known weaknesses and are not considered secure. Additionally, disable the NULL, RC4, DES, and MD5 cipher suites. Ensure all key lengths are greater than 128 bits, use secure renegotiation, and disable compression.
Example: Qualys SSL Labs
Use Valid HTTPS Certificates from a Reputable CAHTTPS certificates should be signed by a reputable certificate authority. The name on the certificate should match the FQDN of the website. The certificate itself should be valid and not expired.
Example: Let's Encrypt (
Disable Data Caching Using Cache Control Headers and AutocompleteBrowser data caching should be disabled using the cache control HTTP headers or meta tags within the HTML page. Additionally, sensitive input fields, such as the login form, should have the autocomplete=off setting in the HTML form to instruct the browser not to cache the credentials.cwe-524
Limit the Use and Storage of Sensitive DataConduct an evaluation to ensure that sensitive data is not being unnecessarily transported or stored. Where possible, use tokenization to reduce data exposure risks.

configuration and operations

Best PracticeDescriptionCWE ID
Automate Application DeploymentAutomating the deployment of your application, using Continuous Integration and Continuous Deployment, helps to ensure that changes are made in a consistent, repeatable manner in all environments.
Establish a Rigorous Change Management ProcessA rigorous change management process must be maintained during change management operations. For example, new releases should only be deployed after process
Example: RBS production outage (
Define Security RequirementsEngage the business owner to define security requirements for the application. This includes items that range from the whitelist validation rules all the way to nonfunctional requirements like the performance of the login function. Defining these requirements up front ensures that security is baked into the system.
Conduct a Design ReviewIntegrating security into the design phase saves money and time. Conduct a risk review with security professionals and threat model the application to identify key risks. The helps you integrate appropriate countermeasures into the design and architecture of the application.cwe-701
Perform Code ReviewsSecurity focused code reviews can be one of the most effective ways to find security bugs. Regularly review your code looking for common issues like SQL Injection and Cross-Site Scripting.cwe-702
Perform Security TestingConduct security testing both during and after development to ensure the application meets security standards. Testing should also be conducted after major releases to ensure vulnerabilities did not get introduced during the update process.
Harden the InfrastructureAll components of infrastructure that support the application should be configured according to security best practices and hardening guidelines. In a typical web application this can include routers, firewalls, network switches, operating systems, web servers, application servers, databases, and application frameworks.cwe-15
Define an Incident Handling PlanAn incident handling plan should be drafted and tested on a regular basis. The contact list of people to involve in a security incident related to the application should be well defined and kept up to date.
Educate the Team on SecurityTraining helps define a common language that the team can use to improve the security of the application. Education should not be confined solely to software developers, testers, and architects. Anyone associated with the development process, such as business analysts and project managers, should all have periodic software security awareness training.


Best PracticeDescriptionCWE ID
Don't Hardcode CredentialsNever allow credentials to be stored directly within the application code. While it can be convenient to test application code with hardcoded credentials during development this significantly increases risk and should be avoided.
Example: Hard coded passwords in networking devices
Develop a Strong Password Reset SystemPassword reset systems are often the weakest link in an application. These systems are often based on the user answering personal questions to establish their identity and in turn reset the password. The system needs to be based on questions that are both hard to guess and brute force. Additionally, any password reset option must not reveal whether or not an account is valid, preventing username harvesting.
Example: Sara Palin password hack (
Implement a Strong Password Policy and Move Towards Passwordless AuthenticationFor password based authentication, password policy should be created and implemented so that passwords meet specific strength criteria. If the user base and application can support it, leverage the various forms of passwordless authentication such as FIDO2 based authentication or mobile application push based authenticators.
Implement Account Lockout against Brute Force AttacksAccount lockout needs to be implemented to guard against brute forcing attacks against both the authentication and password reset functionality. After several tries on a specific user account, the account should be locked for a period of time or until manually unlocked. Additionally, it is best to continue the same failure message indicating that the credentials are incorrect or the account is locked to prevent an attacker from harvesting usernames.cwe-307
Don't Disclose Too Much Information in Error MessagesMessages for authentication errors must be clear and, at the same time, be written so that sensitive information about the system is not disclosed. For example, error messages which reveal that the userid is valid but that the corresponding password is incorrect confirms to an attacker that the account does exist on the system.
Store Database Credentials and API keys SecurelyModern web applications usually consist of multiple layers. The business logic tier (processing of information) often connects to the data tier (database). Connecting to the database, of course, requires authentication. The authentication credentials in the business logic tier must be stored in a centralized location that is locked down. The same applies to accessing APIs providing services to support your application. Scattering credentials throughout the source code is not acceptable. Some development frameworks provide a centralized secure location for storing credentials to the backend database. Secret management solutions that are cloud based or on-premise can be used to allow the application to acquire the credential at application launch or when needed, therefore securing the credentials and avoid storing them statically on disk within a server or a container image.cwe-257
Applications and Middleware Should Run with Minimal PrivilegesIf an application becomes compromised it is important that the application itself and any middleware services be configured to run with minimal privileges. For instance, while the application layer or business layer needs the ability to read and write data to the underlying database, administrative credentials that grant access to other databases or tables should not be provided.cwe-250

Session Management

Best PracticeDescriptionCWE ID
Ensure That Session Identifiers Are Sufficiently RandomSession tokens must be generated by secure random functions and must be of a sufficient length so as to withstand analysis and prediction.cwe-6
Regenerate Session TokensSession tokens should be regenerated when the user authenticates to the application and when the user privilege level changes. Additionally, should the encryption status change, the session token should always be regeneratedcwe-384
Implement an Idle Session TimeoutWhen a user is not active, the application should automatically log the user out. Be aware that Ajax applications may make recurring calls to the application effectively resetting the timeout counter automatically.cwe-613
Implement an Absolute Session TimeoutUsers should be logged out after an extensive amount of time (e.g. 4-8 hours) has passed since they logged in. This helps mitigate the risk of an attacker using a hijacked session.cwe-613
Destroy Sessions at Any Sign of TamperingUnless the application requires multiple simultaneous sessions for a single user, implement features to detect session cloning attempts. Should any sign of session cloning be detected, the session should be destroyed, forcing the real user to re-authenticate.
Invalidate the Session after LogoutWhen the user logs out of the application the session and corresponding data on the server must be destroyed. This ensures that the session can not be accidentally revived.cwe-613
Place a Logout Button on Every PageThe logout button or logout link should be easily accessible to the user on every page after they have authenticated.
Set the Cookie Domain and Path CorrectlyThe cookie domain and path scope should be set to the most restrictive settings for your application. Any wildcard domain scoped cookie must have a good justification for its existence.
Set the Cookie Expiration TimeThe session cookie should have a reasonable expiration time. Non-expiring session cookies should be avoided.

Input and Output Handling

Best PracticeDescriptionCWE ID
Conduct Contextual Output EncodingAll output functions must contextually encode data before sending it to the user. Depending on where the output will end up in the HTML page, the output must be encoded differently. For example, data placed in the URL context must be encoded differently than data placed in JavaScript context within the HTML page.
Example: Resource:
Prefer Whitelists over BlacklistsFor each user input field, there should be validation on the input content. Whitelisting input is the preferred approach. Only accept data that meets a certain criteria. For input that needs more flexibility, blacklisting can also be applied where known bad input patterns or characters are blocked.cwe-159
Use Parameterized SQL QueriesSQL queries should be crafted with user content passed into a bind variable. Queries written this way are safe against SQL injection attacks. SQL queries should not be created dynamically using string concatenation. Similarly, the SQL query string used in a bound or parameterized query should never be dynamically built from user input.
Example: Sony SQL injection Hack (
Use Tokens to Prevent Forged RequestsIn order to prevent Cross-Site Request Forgery attacks, you must embed a random value that is not known to third parties into the HTML form. This CSRF protection token must be unique to each request. This prevents a forged CSRF request from being submitted because the attacker does not know the value of the token.cwe-352
Set the Encoding for Your ApplicationFor every page in your application set the encoding using HTTP headers or meta tags within HTML. This ensures that the encoding of the page is always defined and that browser will not have to determine the encoding on its own. Setting a consistent encoding, like UTF-8, for your application reduces the overall risk of issues like Cross-Site Scripting.cwe-172
Validate Uploaded FilesWhen accepting file uploads from the user make sure to validate the size of the file, the file type, and the file contents as well as ensuring that it is not possible to override the destination path for the file.cwe-434
Use the Nosniff Header for Uploaded ContentWhen hosting user uploaded content which can be viewed by other users, use the X-Content-Type-Options: nosniff header so that browsers do not try to guess the data type. Sometimes the browser can be tricked into displaying the data type incorrectly (e.g. showing a GIF file as HTML). Always let the server or application determine the data type.cwe-430
Validate the Source of InputThe source of the input must be validated. For example, if input is expected from a POST request do not accept the input variable from a GET request.cwe-20
Use the X-Frame-Options Header Use the X-Frame-Options header to prevent content from being loaded by a foreign site in a frame. This mitigates Clickjacking attacks. For older browsers that do not support this header add framebusting Javascript code to mitigate Clickjacking (although this method is not foolproof and can be circumvented).
Example: Flash camera and mic hack (
Use Secure HTTP Response HeadersThe Content Security Policy (CSP), X-XSS-Protection, X-Content-Type-Options headers help defend against Cross-Site Scripting (XSS) attack. In specific, CSP should be customized for the application to lock down the source and location of content plus adding logging to provide some attack detection capability on the front end.
Example: OWASP Secure Headers Project
Deserialize Untrusted Data with Proper ControlsWhen handling serialized data from untrusted source (or passing through untrusted paths), proper controls have to be in place to prevent attacker from abusing the automatic data structure rebuilding capability within the programming language. Each programming platform has its own mitigation strategy which range from using alternative data interchange format such as JSON to restricting the types of objects that can be deserialized. Refer to OWASP Deserialization Cheat Sheet for some great defense information.cwe-502

Access Control

Best PracticeDescriptionCWE ID
Apply Access Controls Checks ConsistentlyAlways apply the principle of complete mediation, forcing all requests through a common security "gate keeper." This ensures that access control checks are triggered whether or not the user is authenticated.cwe-284
Apply the Principle of Least PrivilegeMake use of a Mandatory Access Control system. All access decisions will be based on the principle of least privilege. If not explicitly allowed then access should be denied. Additionally, after an account is created, rights must be specifically added to that account to grant access to resources.cwe-272
Don't Use Direct Object References for Access Control ChecksDo not allow direct references to files or parameters that can be manipulated to grant excessive access. Access control decisions must be based on the authenticated user identity and trusted server side information.cwe-284
Don't Use Unvalidated Forwards or RedirectsAn unvalidated forward can allow an attacker to access private content without authentication. Unvalidated redirects allow an attacker to lure victims into visiting malicious sites. Prevent these from occurring by conducting the appropriate access controls checks before sending the user to the given location.cwe-601