(My) CSSLP Notes – Secure Software Concepts

Note: This notes were strongly inspired by the following book: CSSLP Certification All in one.

General Security Concepts


The security of IT systems can be defined using the following attributes:

  • confidentiality – how the system prevents the disclosure of information.
  • integrity – how the system protects data from the unauthorized access.
  • availability – access to the system by authorized personnel.
  • authentication – process of determining the identity of a user. Three methods can be used to authenticate a user:
    • something you know (ex: password, pin code)
    • something you have (ex: token, card)
    • something you are (ex: biometrics mechanisms)
  • authorization – process of applying access control rules to a user process to determine if a particular user process can access an object.
  • accounting (auditing) – records historical events on a system.
  • non-repudiation – preventing a subject from denying a previous action with an object in a system.

System principles

  • session management – design and implementation of controls to ensure that the communications channels are secured from unauthorized access and disruption of communications.
  • exception management – the process of handling any errors that could appear during the system execution.
  • configuration management – identification and management of the configuration items (initialization parameters, connection strings, paths, keys).

Secure design principles

  • good enough security – there is a trade off between security and other aspects associated with a system. The level of required security must be determined at design time.
  • least privilege – a subject should have only the necessary rights and privileges to perform a specific task.
  • separation of duties – for any given task, more than one individual needs to be involved.
  • defense in depth (layered security) – apply multiple dissimilar security defenses.
  • fail-safe – when a system experience a failure, it should fail to a safe state; all the attributes associated with the system security (confidentiality, integrity, availability) should be appropriately maintained.
  • economy of mechanism – keep the design of the system simple and less complex; reduce the number of dependencies and/or services that the system needs in order to operate.
  • complete mediation – checking permission each time subject requests access to objects.
  • open design – design is not a secret, implementation of safeguard is. (ex: cryptography algorithms are open but the keys used are secret)
  • least common mechanism – minimize the amount of mechanism common to more than one user and depended on by all users. Every shared mechanism (especially one involving shared variables) represents a potential information path between users and must be designed with great care to be sure it does not unintentionally compromise security.
  • psychological acceptability – accessibility to resources should not be inhibited by security mechanisms. If security mechanisms hinder the usability or accessibility of resources, then users may opt to turn off those mechanisms.
  • weakest link – attackers are more likely to attack a weak spot in a software system than to penetrate a heavily fortified component.
  • leverage existing components – component reuse have many advantages, including the increasing of efficiency and security. From the security point of view the component reuse is reducing the attack surface.
  • single point of failure – a system design should not be susceptible to a single point of failure.

Security Models

Access Control Models

Access controls define what actions a subject can perform on specific objects.

  • Bell-LaPadula confidentiality model – It is focused on maintaining the confidentiality of objects. Bell-LaPadula operates by observing two rules: the Simple Security Property and the * Security Property.
    • The Simple security property states that there is “no read up:” a subject at a specific classification level cannot read an object at a higher classification level.
    • The * Security Property is “no write down:”a subject at a higher classification level cannot write to a lower classification level.
  • Take-Grant  – systems specify the rights that a subject can transfer to a from another subject or object. The model is based on representation of the controls in forms of directed graphs with the vertices being the subjects and the objects. The edges between them represent the right between the subject and objects. The representation of rights takes the form of {t (take), g (grant), r (read), w (write)}.
  • Role-based Access control – users are assign a set of roles they may perform. The roles are associated to the access permissions necessary to perform the tasks.
  • MAC (Mandatory Access Control) Model – in MAC systems the owner or subject cannot determine whether access is to be granted to another subject; it is the job of the operating system to decide.
  • DAC (Discretionary Access Control) Model – in DAC systems the owner of an object can decide which other subjects may have access to the object what specific access they may have.

Integrity Models

  • Biba integrity model  – (sometimes referred as Bell-LaPadula upside down) was the first formal integrity model.  Biba is the model of choice when integrity protection is vital. The Biba model has two primary rules: the Simple Integrity Axiom and the * Integrity Axiom. 
    • The Simple Integrity Axiom is “no read down:”a subject at a specific classification level cannot read data at a lower classification. This protects integrity by preventing bad information from moving up from lower integrity levels.
    • The * Integrity Axiom is “no write up:”a subject at a specific classification level cannot write to data at a higher classification. This protects integrity by preventing bad information from moving up to higher integrity levels.
  • Clark-Wilson  –  (this is an informal model) that protects integrity by requiring subjects to access objects via programs. Because the programs have specific limitations to what they can and cannot do to objects, Clark-Wilson effectively limits the capabilities of the subject.Clark-Wilson uses two primary concepts to ensure that security policy is enforced; well-formed transactions and Separation of Duties.

Information Flow Models

Information in a system must be protected when at rest, in transit and in use.

  • The Chinese Wall model – designed to avoid conflicts of interest by prohibiting one person, such as a consultant, from accessing multiple conflict of interest categories (CoIs). The Chinese Wall model requires that CoIs be identified so that once a consultant gains access to one CoI, they cannot read or write to an opposing CoI.


Risk Management


  • risk – possibility of suffering harm or loss
  • residual risk – risk that remains after a control was added to mitigate the initial risk.
  • total risk – the sum of all risks associated with an asset.
  • asset – resource an organization needs to conduct his business.
  • threat – circumstance or event with the potential to cause harm to an asset.
  • vulnerability – any characteristic if an asset that can be exploited by a threat to cause harm.
  • attack – attempting to use a vulnerability.
  • impact – loss resulting when a threat exploits a vulnerability.
  • mitigate – action taken to reduce the likelihood of a threat.
  • control – measure taken to detect, prevent or mitigate the risk associated with a threat.
  • risk assessment – process of identifying risks and mitigating actions.
  • qualitative risk assessment – subjectively determining the impact of an event that effects assets.
  • quantitative risk assessment –  objectively determining the impact of an event that effects assets.
  • single loss expectation (SLE) – linked to the quantitative risk assessment, it represents the monetary loss or impact of each occurrence of a threat.
    • SLE = asset value * exposure factor
  • exposure factor – linked to the quantitative risk assessment, is a measure of the magnitude of a loss.
  • annualized rate of occurrence (ARO) – linked to the quantitative risk assessment, is the frequency with an event is expected to occur on an annualized basis.
    • ARO = number of events / number of years
  • annualized loss of expectancy (ALE) – linked to the quantitative risk assessment, it represents how much an event is expected to cost per year.
    • ALE = SLE * ARO

Types of risks:

  • Business Risks:
    • fraud
    • regulatory
    • treasury management
    • revenue management
    • contract management
  • Technology Risks:
    • security
    • privacy
    • change management

Types of controls

Controls can be classified on types of actions they perform. Three classes of controls exist:

  • administrative
  • technical
  • physical

For each of these classes, there are four types of controls:

  • preventive (deterrent) – used to prevent the vulnerability
  • detective – used to detect the presence of an attack.
  • corrective (recovery) – correct a system after a vulnerability is exploited and an impact has occurred; backups are  a common form of corrective controls.
  • compensation – designed to act when a primary set of controls has failed.

Risk management models

General risk management model

The steps contained in a general risk management model:

  1. Asset identification – identify and clarify all the assets, systems and processes that need to be protected.
  2. Threat assessment – identify the threats and vulnerabilities associated with each asset.
  3. Impact determination and qualification
  4. Control design and evaluation – determine which controls to put in place to mitigate the risks.
  5. Residual risk management – evaluate residual risks to identify where additional controls are needed.

Risk management model proposed by Software Engineering Institute

SEI model steps :

  1. Identity – enumerate potential risks.
  2. Analyze – convert the risk data gather into information that can be used to make decisions.
  3. Plan – decide the actions to take to mitigate them.
  4. Track – monitor the risks and mitigations plans.
  5. Control – make corrections for deviations from the risk mitigation plan.

Security Policies and Regulations

One of the most difficult aspects of prosecution of computer crimes is attribution. Meeting the burden of proof requirement in criminal proceedings, beyond a reasonable doubt, can be difficult given an attacker can often spoof the source of the crime or can leverage different systems under someone else’s control.

Intellectual property

Intellectual property is protected by the U.S law under one of four classifications:

  • patents – Patents provide a monopoly to the patent holder on the right to use, make, or sell an invention for a period of time in exchange for the patent holder’s making the invention public.
  • trademarks – Trademarks are associated with marketing: the purpose is to allow for the creation of a brand that distinguishes the source of products or services.
  • copyrights – represents a type of intellectual property that protects the form of expression in artistic, musical, or literary works, and is typically denoted by the circle c symbol. Software is typically covered by copyright as if it were a literary work. Two important limitations on the exclusivity of the copyright holder’s monopoly exist: the doctrines of first sale and fair use. The first sale doctrine allows a legitimate purchaser of copyrighted material to sell it to another person. If the purchasers of a CD later decide that they no longer cared to own the CD, the first sale doctrine gives them the legal right to sell the copyrighted material even though they are not the copyright holders.
  • trade secrets – business-proprietary information that is important to an organization’s ability to compete. Software source code or firmware code are examples of computer-related objects that an organization may protect as trade secrets.

Privacy and data protection laws

Privacy and data protection laws are enacted to protect information collected and maintained on individuals from unauthorized disclosure or misuse.

Several important pieces of privacy and data protection legislation include :

  • U.S. Federal Privacy Act of 1974 – protects records and information maintained by U.S. government agencies about U.S. citizens and lawful permanent residents.
  •  U.S. Health Insurance Portability and Accountability Act (HIPAA) of 1996 – seeks to guard protected health information from unauthorized use or disclosure.
  • Payment Card Industry Data Security Standard (PCI-DSS) – the goal is to ensure better protection of card holder data through mandating security policy, security devices, control techniques and monitoring of systems and networks.
  • U.S. Gramm-Lech-Bliley Financial Services Modernization Act (GLBA) – requires financial institutions to protect the confidentiality and integrity of consumer financial information.
  • U.S. Sarbanes-Oxley Act of 2002 (SOX) – the primary goal of SOX is to ensure adequate financial disclosure and financial auditor independence.

Secure Software Architecture – Security Frameworks

  • COBIT (Control Objectives for Information and Related Technology)– assist management in bringing the gap between control requirements, technological issues and business risks.
  • COSO (Committee of Sponsoring Organizations of the Treadway Commission) – COSO has established a Enterprise Risk Management -Integrated Framework against which companies and organizations may assess their control systems.
  • ITIL (Information Technology Infrastructure Library) – describes a set of practices focusing on aligning IT services with business needs.
  • SABSA (Sherwood Applied Business Security Architecture) – framework and methodology for developing risk-driven enterprise information security architecture.
  • CMMI (Capability Maturity Model Integration) – process metric model that rates the process maturity of an organization on a 1 to 5 scale.
  • OCTAVE (Operationally Critical Threat, Asset and Vulnerability Evaluation) – suite of tools, techniques and methods for risk-based information security assessment.


Software Development Methodologies

Secure Development Lifecycle Components

  • software team awareness and education – all team members should have appropriate training. The key element of team awareness and education is to ensure that all the members are properly equipped with the correct knowledge.
  • gates and security requirements – the term gates it signifies a condition that one must pass through. To pass the security gate a review of the appropriate security requirements is conducted.
  • threat modeling – design technique used to communicate information associated with a threat throughout the development team (for more infos’ you could check my other ticket : threat modeling for mere mortals).
  • fuzzing – a test technique where the tester applies a series of inputs to an interface in an automated fashion and examines the output for undesired behaviors.
  • security reviews – process to ensure that the security-related steps are being carried out and not being short-circuited.

Software Development Models

  • waterfall model – is a linear application development model that uses rigid phases; when one phase ends, the next begins.
  • spiral model – repeats steps of a project, starting with modest goals, and expanding outwards in ever wider spirals (called rounds). Each round of the spiral constitutes a project, and each round may follow traditional software development methodology such as Modified Waterfall. A risk analysis is performed each round.
  • prototype model – working model of software with some limited functionality. Prototyping is used to allow the users evaluate developer proposals and try them out before implementation.
  • agile model
    • Scrum  – contain small teams of developers, called the Scrum Team. They are supported by a Scrum Master, a senior member of the organization who acts like a coach for the team. Finally, the Product Owner is the voice of the business unit.
    • Extreme Programming (XP) – method that uses pairs of programmers who work off a detailed specification.

Microsoft Security Development Lifecycle

SDL is software development process designed ti enable development teams to build more secure software and address security compliance requirements.

SDC is build around the following three elements:

  • (security) by design – the security thinking is incorporated as part of design process.
  • (security) by default – the default configuration of the software is by design as secure as possible.
  • (security) in deployment – security and privacy elements are properly understood and managed through the deployment process.

SDL components:

  • training   security training for all personnel, targeted to their responsibility associated with the development effort.
  • requirements
    • establishment of the security and privacy requirements for the software.
    • creation of quality gates ans bug bars. Defining minimum acceptable levels of security and privacy quality at the start helps a team understand risks associated with security issues, identify and fix security bugs during development, and apply the standards throughout the entire project.Setting a meaningful bug bar involves clearly defining the severity thresholds of security vulnerabilities (for example, no known vulnerabilities in the application with a “critical” or “important” rating at time of release) and never relaxing it once it’s been set.
    • development of security and privacy risk assessment. Examining software design based on costs and regulatory requirements helps a team identify which portions of a project will require threat modeling and security design reviews before release and determine the Privacy Impact Rating of a feature, product, or service.
  • design – establish design requirements, perform attack/surface analysis/reduction and use threat modeling.
  • implementation – application of secure coding practices and the use of static program checkers to find common errors.
  • verification – perform dynamic analysis (tools that monitor application behavior for memory corruption, user privilege issues, and other critical security problems), fuzz testing and conduct attack surface review.
  • release – conduct final security review and create an incident response plan.
  • response – execute incident response plan.

How to remotely connect to an in-memory HSQLDB database

hsqldbVery often in-memory instances of HSQLDB are used in the context of unit tests; the unit test starts a database instance (eventually on a random port), provision the database with some data, run the test against the database end then stop it.

Now, from time to time you need to debug the unit tests and sometimes you also need to run manually some queries on the in-memory database (using HSQLDB manager or any other software).

So here are the steps in order to be able to connect to an in-memory HSQLDB instance:

  • Start the DB in the “remote open db” mode.This is driven by the “server.remote_open” property that is false by default (you can look to the code of org.hsqldb.server.ServerProperties class to see other properties that might be interesting). The code that starts a dataabse instacne in remote open mode will look something like:
HsqlProperties props = new HsqlProperties();
props.setProperty("server.remote_open", true);
....add more properties here
Server server = new Server();
  • Connect to the data base instance. Now that the database has started the next step is to connect to data base instance. Because we started an in-memory instance the connection url is a memory database url that looks like jdbc:hsqldb:mem:instanceName You will not be able to connect using this url because neither the host and the port are available, instead you should use a server database url that looks like jdbc:hsqldb:hsql://host:port/instanceName


How to find (buggy) calls of java.lang.Object.equals() with incompatible types


All this started from a change (and a mistake) done by one of my co-workers that did a small change in the code with rather dramatic consequences;

So, imagine we have a (java) class T having a private field f of type String:

public class T {
    private Boolean f;

    public void setF(final Boolean f) {
        this.f = f;
    public Boolean getF() {
        return f;

Now, imagine (again) that the type of f field will change from String to something else, let’s say Boolean; the compiler will help you to find the places where the setF method is called with the wrong parameter and also where the getF method do not comply with the new signature.

But there is at least one case where the compiler cannot help you; it is the case of the call of java.lang.Object.equals method. So something like “aString”.equals(t.getF()) in the case when the f field is a boolean will always return false because we try to check the equality for incompatible types.

The goal of this ticket is to find solutions to catch this kind of problem at runtime or (ideally) at compile time. Anyway, if you want to catch this kind of problem in your code and you do not want to reinvent the wheel (as I will do in this ticket), just use the FindBugs. FindBugs is capable to catch this kind of error (EC_UNRELATED_TYPES) and many, many more (see FindBugs Bug Descriptions).

How to find the buggy calls at runtime

The only way that I see to catch the buggy equals calls in a generic way is to use AOP. The basic idea is to write an aspect that will intercept all the calls to the equals methods and check that the caller and the callee are of the same type.

For the implementation I choose to use AspectJ framework (for a very good introduction to AspectJ I strongly recommend the AspectJ in action book) and the code it looks like this:

import org.aspectj.lang.annotation.Aspect;
import org.aspectj.lang.annotation.Before;
import org.aspectj.lang.annotation.Pointcut;
import org.github.adriancitu.equals.WrongUseOfEqualsException;

public class EqualsCheckerAspect {
            "execution(public boolean Object+.equals(Object)) "
            + "&& target(target) "
            + "&& args(methodArgument)"
            + "&& !within(EqualsCheckerAspect)")
    public void executeEqualsPointcut(
            final Object target, 
            final Object methodArgument) {
    public void executeEqualsAdvice(
            final Object target, 
            final Object methodArgument) throws Exception {
        if (target != null && !target.getClass().equals(methodArgument.getClass())) {
            throw new WrongUseOfEqualsException("Tried to call the equals methods on different classes types");

I will try briefly to explain what the previous code is doing; this is not a tutorial about AspectJ or AOP so, some of the definitions will be a little bit approximative, so AOP purists please forgive me.

The aspect (which is a Java class annotated with the @Aspect annotation) contains 2 parts; the pointcut (the method annotated with the @Pointcut annotation) and the advice (the method annotated with the @Before annotation).

The pointcut represents the parts of program flow that we want to intercept; in our case we want to intercept all the calls to the method having the following signature boolean equals(Object) on any instance of java.lang.Object class and on any subclasses (that’s the meaning of the + from the ..boolean Object+.equals). The  target and args are just AspectJ structures that helps to capture the caller and the callee of the equals method.

The advice represents the code that will be executed when a poincut is intercepted. In our  case we would like to do the verification that the caller and the callee are of the same type before the execution of the equals method (this is the meaning of the @Before annotation). The advice code represented by the executeEqualsAdvice method is quite simple, it just retrieve the caller and the callee objects (already captured by the target and args) and check that are instances of the same class.

This solution will fix the problem but it have some drawbacks:

  • the check is done at runtime
  • you have a dependency on AspectJ

How to find the buggy calls at compile time

It would be really nice to be able to catch this problem directly at compile time and eventually (not mandatory) having no other technical dependency.

Our salvation is coming from the javac plug-in mechanism which is new in Java8. The javac plug-in mechanism allows to a user to specify one or more plug-ins on the javac command line, to be started soon after the compilation has begun. There is no official tutorial about how to craft and use a javac plug-in, the only information that I’ve found is this Javadoc link and this blog ticket.

The javac plug-in mechanism gives access to the abstract syntax tree of the compiled program by implementing the visitor pattern so we will use this mechanism to catch and check the executions of the equals methods. The entire code of the plug-in can be found on this GitHub repository.

The steps of crafting a (javac) plug-in are the following ones:

  1. The entry point of the plug-in should implement the com.sun.source.util.Plugin interface. The code of the class can be found here : RuntimeEqualsCheckPlugin.java
  2. Create a class that implements the com.sun.source.util.TaskListener interface in order to perform additional behavior after the type-checking phase. The code of the class can be found here: EqualsMethodCallTaskListener.java
  3. Create an abstract syntax tree visitor by extending the com.sun.source.util.TreePathScanner<R,P>
    end extend the behavior for the method invocation. The code for this class can be found here: CodePatternTreeVisitor2.java

The most important part of CodePatternTreeVisitor2.java class is the overridden visitMethodInvocation method. I will show you the code but the most important lesson that I learned is that in the code you have to work with trees, everything is a subtype of the Tree interface.

public Void visitMethodInvocation(MethodInvocationTree methodInvocationTree, Void aVoid) {
  final List<? extends ExpressionTree> arguments = methodInvocationTree.getArguments();
  final ExpressionTree methodSelect = methodInvocationTree.getMethodSelect();
  switch (methodInvocationTree.getKind()) {
        Tree.Kind methodSelectKind = methodSelect.getKind();
        switch (methodSelect.getKind()) {
            case MEMBER_SELECT:
             MemberSelectTree memberSelectTree = (MemberSelectTree) methodSelect;
             //it's a equals method invocation
             if (isEqualsCall(
                  new TreePath(getCurrentPath(), methodSelect),
                  methodInvocationTree.getArguments() != null ?
                       methodInvocationTree.getArguments().size() : 0)) {
                ExpressionTree expression = memberSelectTree.getExpression();
                TypeMirror callerType =
                            new TreePath(getCurrentPath(), expression));
                Optional argumentType =
                if (argumentType.isPresent() 
                      && !callerType.equals(argumentType.get())) {
                      System.err.println("Try to call equals on different parameters at line "
                          + getLineNumber(methodInvocationTree)
                          + " of file " +
                          + "; this is a bug!"
        return super.visitMethodInvocation(methodInvocationTree, aVoid);

The steps to execute the (javac) plug-in are the following ones:

  1. Set up a file called com.sun.source.util.Plugin located in META-INF/services/ of your plug-in code because the plug-ins is located via java.util.ServiceLoader. The content of the file should contain the name of the plug-in (com.github.adriancitu.equals.com.github.adriancitu.equals.runtime.RuntimeEqualsCheckPlugin in this case).
  2. Create a jar with all the plug-in files and the com.sun.source.util.Plugin file.
  3. Execute your plug-in against the javac using the  -processorpath flag to indicates the path where the plug-in JAR file is located and -Xplugin flag to indicate the name of the plug-in to run. In our case the command will be something like :
    javac -processorpath plugIn.jar
    -Xplugin:RuntimeEqualsCheckPlugin ./NoDependenciesTest.java

So finally was possible to have a compile time solution, the only drawback that I see for the plug-in solution is that the API is not very user friendly.

A Java implementation of CSRF mitigation using “double submit cookie” pattern

Goal of this articlecsrf

The goal of this article is to present an implementation of the “double submit cookie” pattern used to mitigate the Cross Site Request Forgery (CSRF) attacks. The proposed implementation is a Java filter plus a few auxiliary classes and it is (obviously) suitable for projects using the Java language as back-end technology.

Definition of CSRF and possible mitigations

In the case of a CSRF attack, the browser is tricked into making unauthorized requests on the victim’s behalf, without the victim’s knowledge. The general attack scenario contains the following steps:

  1. the victim connects to the vulnerable web-site, so it have a real, authenticated session.
  2. the hacker force the victim (usually using a spam/fishing email) to navigate to another (evil) web-site containing the CSRF attack.
  3. when the victim browser execute the (evil) web-site page, the browser will execute a (fraudulent) request to the vulnerable web-site using the user authenticated session. The user is not aware at all of the fact that navigating on the (evil) web-site will trigger an action on the vulnerable web-site.

For deeper explanations I strongly recommend  to read chapter 5 of Iron-Clad Java: Building Secure Applications book and/or the OWASP Cross-Site Request Forgery (CSRF) Prevention Cheat Sheet.

Definition of “double submit cookie” pattern

When a user authenticates to a site, the site should generate a (cryptographically strong) pseudo-random value and set it as a cookie on the user’s machine separate from the session id. The server does not have to save this value in any way, that’s why this patterns is sometimes also called Stateless CSRF Defense.

The site then requires that every transaction request include this random value as a hidden form value (or other request parameter). A cross origin attacker cannot read any data sent from the server or modify cookie values, per the same-origin policy.

In the case of this mitigation technique the job of the client is very simple; just retrieve the CSRF cookie from the response and add it into a special header to all the requests:


Client workflow

The job of the server is a little more complex; create the CSRF cookie and for each request asking for a protected resource, check that the CSRF cookie and the CSRF header of the request are matching:


Server workflow

Note that some JavaScript frameworks like AngularJS implements the client worflow out of the box; see Cross Site Request Forgery (XSRF) Protection

Java implementation of “double submit cookie” pattern

The proposed implementation is on the form of a (Java) Servlet filter and can be found here: GenericCSRFFilter GitHub.

In order to use the filter, you must define it into you web.xml file:




The filter can have 2 optional initialization parameters: csrfCookieName representing the name of the cookie that will store the CSRF token and csrfHeaderName representing the name of the HTTP header that will be also contains the CSRF token.

The default values for these parameters are “XSRF-TOKEN” for the csrfCookieName and “X-XSRF-TOKEN” for the csrhHeaderName, both of them being the default values that AngularJS is expecting to have in order to implement the CSRF protection.

By default the filter have the following features:

  • works with AngularJS.
  • the CSRF token will be a random UUID.
  • all the resources that are NOT accessed through a GET request method will be CSRF protected.
  • the CSRF cookie is replaced after each non GET request method.

How it’s working under the hood

The most of the functionality is in the GenericCSRFStatelessFilter#doFilter method; here is the sequence diagram that explains what’s happening in this method:

doFilter method sequence diagram

doFilter method sequence diagram

The doFilter method is executed on each HTTP request:

  1. The filter creates an instance of ExecutionContext class; this class is a simple POJO containing the initial HTTP request, the HTTP response, the CSRF cookies (if more than one cookie with the csrfCookieName is present) and implementation of the ResourceCheckerHook , TokenBuilderHook and ResponseBuilderHook .(see the next section for the meaning of this classes).
  2. The filter check the status of the HTTP resource, that status can be:MUST_NOT_BE_PROTECTED, MUST_BE_PROTECTED_BUT_NO_COOKIE_ATTACHED,MUST_BE_PROTECTED_AND_COOKIE_ATTACHED (see ResourceStatus enum) using an instance of ResourceCheckerHook.
  3. If the resource status is ResourceStatus#MUST_NOT_BE_PROTECTED
    the filter creates a CSRF cookie having as token the token generated by an instance of TokenBuilderHook.
  4. if the resource status ResourceStatus#MUST_BE_PROTECTED_AND_COOKIE_ATTACHED
    then compute the CSRFStatus of the resource and then use an instance of ResponseBuilderHook to return the response to the client.

How to extend the default behavior

It is possible to extend or overwrite the default behavior by implementing the hooks interfaces. All the hooks implementations must be thread safe.

  1. The ResourceCheckerHook is used to check the status of a requested resource. The default implementation is DefaultResourceCheckerHookImpl and it will return ResourceStatus#MUST_NOT_BE_PROTECTED for any HTTP GET method, for all the other request types, it will return [email protected] ResourceStatus#MUST_BE_PROTECTED_BUT_NO_COOKIE_ATTACHED if any CSRF cookie is present in the query or ResourceStatus#MUST_BE_PROTECTED_BUT_NO_COOKIE_ATTACHED otherwise.The interface signature is the following one:
    public interface ResourceCheckerHook extends Closeable {
        ResourceStatus checkResourceStatus(ExecutionContext executionContext);
  2. The TokenBuilderHook hook is used to generate the token that will be used to create the CSRF cookie. The default implementation  is DefaultTokenBuilderHookImpl and it uses a call to UUID.randomUUID to fetch a token. The interface signature is the following one:
    public interface TokenBuilderHook extends Closeable {
        String buildToken(ExecutionContext executionContext);
  3. The ResponseBuilderHook is used to generate the response to the client depending of the CSRFStatus of the resource. The default implementation is DefaultResponseBuilderHookImpl and it throws a SecurityException if the CSRF status is CSRFStatus#COOKIE_NOT_PRESENT, CSRFStatus#HEADER_TOKEN_NOT_PRESENT or CSRFStatus#COOKIE_TOKEN_AND_HEADER_TOKEN_MISMATCH. If the CSRF status is CSRFStatus#COOKIE_TOKEN_AND_HEADER_TOKEN_MATCH then the old CSRF cookies are deleted and a new CSRF cookie is created. The interface signature is the following one:
    public interface ResponseBuilderHook extends Closeable {
        ServletResponse buildResponse(ExecutionContext executionContext,
                                      CSRFStatus status);

The hooks are instantiated inside the GenericCSRFStatelessFilter#init method using the ServiceLoader Java 6 loading facility. So if you want to use your implementation of one of the hooks then you have to create a  META-INF/services directory that contains a text file whose name matches the fully-qualified interface class name of the hook that you want to replace.

Here is the sequence diagram representing the hooks initializations:


Book review : Practical Anonymity: Hiding in Plain Sight Online

This is a review of the Practical Anonymity: Hiding in Plain Sight Online book.


This is not a technical book about the inner workings of Tor or Tails and I think a better title would be “How to use Tor and Tails for dummies”. Almost all the information present in the book can be found in the official documentation, the only positive point is that all the needed information is present in one single place.

Chapter 1. Anonymity and Censorship Circumvention

This first chapter is an introduction to what is on-line anonymity, why (the on-line anonymity ) is important for some people and how it can be achieved using Tor. The chapter contains also the fundamentals of how Tor is working, what it can do to on-line anonymity and some advises about how it can be used safely.

Chapter 2. Using the Tor Browser Bundle

The chapter is presenting the TBB (Tor Browser Bundle) in detail. The TBB is composed of three components; Vidalia which is the control panel for Tor, the Tor software itself and Mozilla Firefox browser. Each of this tree components are described from the user point of view, each of the possible configuration options are presented in detail.

Chapter 3. Using Tails

Tails is a is a Linux distribution that includes Tor and other softwares to provide an operating system that enhances privacy. The Tails network stack has been modified so that all Internet connectivity is routed through the Tor network.

In order to enhance privacy, Tails is delivered with the following packages :

  • Firefox
  • Pidgin
  • GNU Privacy Guard
  • Metadata Anonymization Toolkit
  • Unsafe Web Browser

Detailed instructions are presented about how to create a bootable DVD and a bootable USB stick and how to run and configure the operating system. The persistent storage feature of Tails is presented in detail so that the reader can understand what are the benefits and the drawbacks.

Chapter 4. Tor Relays, Bridges and Obfsproxy

The chapter is about how the Tor adversaries can disrupt the network and how the Tor developers are trying to find new technique to workaround these disruptions.

One way to forbid to the user the access to the Tor network is to filter the (nine) Tor directory authorities, that are servers that distribute information about active Tor entry points. One way to avoid this restriction is the use of Tor bridge relays. A bridge relay is like any other Tor transit relay, the  only difference is that it is not publicly listed and it is used only for entering the Tor network from places where public Tor relays are blocked. There are different mechanisms to retrieve the list of this bridge relays, like a web page on the Tor website or emails sent by email.

Another way to disrupt the Tor network is to filter the Tor traffic knowing that the Tor protocol packages have a distinguished signature. One way to avoid the package filtering is to conceal the Tor packages in  other kind of packages. The framework that can be used to implement this kind of functionality is called Obfsproxy (obfuscated proxy). Some of the plug-ins that are using Pbfsproxy: StegoTorus, Dust, SkypeMorph.

Chapter 5. Sharing Tor Resources

This chapter describes how a user can share his bandwidth becoming a Tor bridge relay, a Tor transit relay or a Tor exit relay. Detailed settings descriptions are made for each type of relays and also the incurred risks for the user.

Chapter 6. Tor Hidden Services

A (Tor) hidden service is a server that can be accessed by other clients within the anonymity network, while the actual location (IP address) of the server remains anonymous. The hidden service protocol is briefly presented followed by how to set up a hidden service. For the set up a hidden service the main takeaways are:

  • install Tor and the service that you need on a VM.
  • run the VM on a VPS (virtoual private server)  hosted  in a country having privacy-friendly legislation in place.
  • the VM can/should be encrypted, be power cycled and that has no way to know what IP address or domain name of the computer on which it is running.

Chapter 7. Email Security and Anonymity practices.

This last chapter is about the email anonymity in general and how the use of Tor can improve the email anonymity. The main takeaways :

  • choose a email provider that do not require another email address or a mobile phone.
  • choose an email provider that supports HTTPS.
  • encrypt the content of your emails.
  • register and connect to the email box using ALWAYS Tor.



(My) BruCON 2016 notes (3)

Here are my quick notes from the BruCON 2016 conference. All the slides can be found here.

NO EASY BREACH:Challenges and Lessons Learned from an Epic Investigation bruCon

The attack started with a phishing email; the attack compromised more that 2 000 systems, 50 000 emails.

How the attack took place:

1. fast-paced attacher

  • 10-25 systems infected every day.
  • the attacker steal information every day.


  • develop indicators to aid triage.
  • focus on : lateral movements, pivoting, recon, new tools or back-doors.
  • streamlined documentation.

lessons learned

  • be fast and flexible.

2. stealthy attacks

  • used anti-forensics techniques to hide endpoint and network activity.
  • altered communication scheme + strong crypto.
  • mass activity to obscure the real target.
  • data theft using only legitimate us-based services – gmail, google drives, one drive.


  • maximize the utility of trace forensics artifacts.
  • some attacker behavior recovered from sdelete.
  • took time and patience to filter out the network noise.
  • deployed additional open-source tech

lessons learned

  • improve visibility and don’t stop looking.
  • map attacker activity ti potential data sources.
  • network times provides reliable chronology.

3. rapidly evolving tactics

  • seven unique persistence mechanism.
  • seven distinct back-door families.
  • minimal re-use of meta-data commonly tracked and shared as indicator.


  •  fought to keep network visibility on all malware families.
  • spent time analyzing system with unknown activity.
  • create indicators for every stage of attack life-cycle.
  • develop flexible & resilient indicators

lessons learned

  • enhance and test your best indicators even when they’re working.

4. advanced attack techniques

  • attacker leveraged PowerShell.
  • used Windows Management Instrumentation.
  • attacker used Kerberos tickets attacks which made tracking lateral movement difficult.


  • searched for WMI persistence.
  • identified evidence of attacker code in WMI repository.
  • parsed out embedded scripts and malware.
  • updated the environment to power shell 3.0 and enabled logging.
  • turned attacker power shell usage from a threat to a benefit by logging and iocs to made findings attacker activity much easier.
  • worked around Kerberos attacks: looked for remote Kerberos logons around the time of attacker activity.

Hacking KPN: Lessons from the trenches

The presentation was about 3 different vulnerabilities discovered by the kpn read team.

  1. vulnerability linked to the Java de-serialisation vulnerability.
    1. the kpn team did a java deserialization burp plug-in fork
  2. Citrix Netscaler
    1. Netscaler login vulnerabiilty
  3. reverse-engineering cryptography from binary

New Adventures in Active Defense, Offensive Countermeasures and Hacking Back

The idea was that the security industry are doing the same things over and over again, very often as a defender we build very static walls. So the presenter propose to an “active defense”:

Active defense is not about :

  • hacking back
  • about one technical solution
  • revenge

Active defense is about:

  • have a range of solutions.

All the proposed solutions and demos are part of the advanced defense harbinger distribution which is a Linux distribution based on Ubuntu LTS that it comes with many tools aimed at active defense pre-installed and configured. Some demos of the following components:

  • weblabyrinth
  • honey ports
  • honey badger
  • jarcombiner

(My) BruCON 2016 notes (2)

Here are my quick notes from the BruCON 2016 conference. All the slides can be found here.

What Does the Perfect Door or Padlock Look Like?bruCon

The talk was about how (some) doors and padlocks can be easily opened. The presentation was full of videos and explanatory schemes. For the doors the following parts can be attacked:

  • hinge removal – to fix, use jam pins
  • the latch
  • the inside handle
  • key boxes
  • edge baps – request to exit sensors
  • the bottom gap
  • the door frame

Anti-Forensics AF

The presentation contained the following topics:

  • memory anti-forencics
    • the goal is to inhibit the acquisition and analysis
    • for windows, removing PE header from disk (once the executable is loaded in memory).
    • for windows, zero the header from disk (once the executable is loaded in memory).
    • for linux, remove the EMF header
    • for linux, zero the header (memeset)
  • android anti-forestics
    • use encryption to protect.
    • power down the device.
    • leverage device sensors; foe ex: if the phone is moving, then shut down the device
  • fun with sd cards
    • demo of the SDTool tool that modifies the SD card firmware to write/or not the card or in memory.

Esoteric Web application vulnerabilities

The sql injection vulnerability is dead due to the massive use of the ORM frameworks, the same for the XSS injections due to the mvc, templates and default HTML So, as a hacker you must find new vulnerabilities; here are 5 (esoteric) vulnerabilities:

  1. aggressive input decoding; nosql injection using ruby on rails and MongoDB
  2. call me to verify my identity; try to hack the phone activation procedure for a 2 FA functionality.
  3. password reset implementation feature; try to hack the password reset feature for a 2 FA
  4. hack around the usage of the Paypal IPN protoco
  5. just missed that one; it happen even to the best of us.

Scraping Leaky browsers for Fun and Password

The idea is to retrieve passwords stored by the browsers (in RAM) by scrapping the RAM content. Do to this a plug-in to Volatility framework was created (the plug-in will be available soon on GitHub).

The best browser is Chrome with 67% chances to expose the passwords; the worst browser is Firefox with 81% changes to expose the passwords.

Vendors response to this findings; Microsoft created a CVE and the path will be pushed in October/November, Google and Mozilla are denying that’s a real issue.