Application Security

The GlobaLeaks software tries to conform with industry standard best and practices and its security is a result of applied research.

This document tries to detail every aspect implemented by the application in relation to the security design.

Architecture

The software is made up of two main components: a Backend and a Client:

  • The Backend is a python backend that runs on a physical backend and exposes a REST API.

  • The Client is a JavaScript client-side web application that interacts with Backend only through XHR.

Anonymity

Users’s anonymity is protected by means of the Tor technology.

The entire application considers to avoid logging of sensible metadata that could lead to identification of whistleblowers.

Authentication

The confidentiality of the authentication is protected by either Tor Onion Services v3 or TLS version 1.2+.

This section describes the authentication methods implemented by the system.

Password

By accessing the login web interface, Administrators and Recipients need to insert their respective Username and Password. If the password submitted is valid, the system grants access to the functionality available to that user.

Receipt

Whistleblowers access their Reports by using a anonymous Receipts, which are random generated 16 digits sequences created by the Backend when the Report is first submitted. The reason of this format of 16 digits is that it resembles a standard phone number, making it easier for the whistleblowers to conceal their receipts.

Password Security

The following password security measures implemented by the system.

Password Storage

Password are never stored in plaintext but the system maintain at rest only an hash. This apply to every authentication secret included whistleblower receipts.

The platform stores Users’ passwords hashed with a random 128 bit salt, unique for each user.

Passwords are hashed using Argon2, a key derivation function that was selected as the winner of the Password Hashing Competition in July 2015.

The hash involves a per-user salt for each user and a per-system salt for whistleblowers.

Password Complexity

The system enforces the usage of complex password by implementing a custom algorithm necessary for ensuring a reasonable entropy of each authentication secret.

Password are scored in three levels: Strong, Acceptable, Insecure.

  • Strong: A strong password should be formed by capital letters, lowercase letters, numbers and a symbols, be at least 12 characters long and include a variety of at least 10 different inputs.

  • Acceptable: An acceptable password should be formed by at least 3 different inputs over capital letters, lowercase letters, numbers and a symbols, be at least 10 characters and include a variety of at least 7 different inputs.

  • Insecure: A password ranked below the strong or acceptable levels is marked as insecure and not accepted by the system.

We encourage each end user to use KeePassXC to generate and retain strong and unique passphrases.

Two Factor Authentication

The system implements Two Factor Authentication (2FA) based on TOTP based on RFC 6238 algorithm and 160 bits secrets.

Users are enabled to enroll for 2FA via their own preferences and administrators can optionally enforce this requirement.

We recommend using FreeOTP available for Android and for iOS.

Password Change on First Login

The system enforces users to change their own password at their first login.

Administrators could as well enforce password change for users at their next login.

Periodic Password Change

By default the system enforces users to change their own password at least every year.

This period is configurable by administrators.

Proof of Work on Login and Submissions

The system implements an automatic Proof of Work on every login that requires every client to request a token, solve a computational problem before being able to perform a login or file a submission.

Rate Limit on Anonymous Sessions

The system implements rate limiting on whistleblowers’ sessions preventing to execute more than 5 requests per second.

Slowdown on Failed Login Attempts

The system identifies multiple failed login attempts and implement a slowdown procedure where an authenticating client should wait up to 42 seconds to complete an authentication.

This feature is intended to slow down possible attacks requiring more resources to users in terms of time, computation and memory.

Password Recovery

In case of password loss users could request a password reset via the web login interface clicking on a Forgot password? button present on the login interface.

When this button is clicked, users are invited to enter their username or an email. If the provided username or the email correspond to an existing user, the system will provide a reset link to the configured email.

By clicking the link received by email the user is then invited to configure a new email different from the previous.

In case encryption is enabled on the system, a user clicking on the reset link would have first to insert their Account Recovery Key and only in case of correct insertion the user will be enabled to set a new password.

In case 2FA is enabled on the system, a user clicking on the reset link would have first to insert an authentication code taken from the authentication API.

Web Application Security

This section describes the Web Application Security implemented by the software in adherence with the OWASP Security Guidelines.

Session management

The session implementation follows the OWASP Session Management Cheat Sheet security guidelines.

The system assigns a Session to each authenticated user. The Session ID is 256bits long secret generated randomly by the backend. Each session expire accordingly to a timeout of 60 minutes. Session IDs are exchanged by the client with the backend by means of an header (X-Session) and do expire as soon that users close their browser or the tab running GlobaLeaks. Users could explicitly log out via a logout button or implicitly by closing the browser.

Cookies and XSRF Prevention

Cookies are not used intentionally to minimize XSRF attacks and any possible attack based on them. Instead than using Cookies authentication is based on a custom HTTP Session Header sent by the client on authenticated requests.

HTTP Headers

The system implements a large set of HTTP headers specifically configured to improve the software security and achieves score A+ by Security Headers and score A+ by Mozilla Observatory.

Strict-Transport-Security

The system implements strict transport security by default.

Strict-Transport-Security: max-age=31536000; includeSubDomains

Administrators could optionally enable the preload functionality.

Content-Security-Policy

The backend implements the following Content Security Policy (CSP):

Content-Security-Policy: default-src 'none'; script-src 'self'; connect-src 'self'; style-src 'self'; img-src 'self' data:; font-src 'self' data:; media-src 'self'; form-action 'self'; frame-ancestors 'none'

Cross-Origin-Embedder-Policy

The backend implements the following Cross-Origin-Embedder-Policy (COEP):

Cross-Origin-Embedder-Policy: require-corp

Cross-Origin-Opener-Policy

The backend implements the following Cross-Origin-Opener-Policy (COOP):

Cross-Origin-Resource-Policy: same-site

Cross-Origin-Resource-Policy

The backend implements the following Cross-Origin-Resource-Policy (CORP):

Cross-Origin-Resource-Policy: same-site

Permissions-Policy

The backend implements the following Permissions-Policy header configuration to limit the possible de-anonymization of the user by disabling dangerous browser features:

Permissions-Policy: camera=('none') display-capture=('none') document-domain=('none') fullscreen=('none') geolocation=('none') microphone=('none') speaker=('none')

X-Frame-Options

The backend configure the X-Frame-Options header to prevent inclusion by means of Iframes in any site:

X-Frame-Options: deny

Referrer-Policy

Web-browsers usually attach referrers in their http headers as they browse links. The platform enforce a referrer policy to avoid this behaviour.

Referrer-Policy: no-referrer

X-Content-Type-Options

When setting up Content-Type for the specific output, we avoid the automatic mime detection logic of the browser by setting up the following header:

X-Content-Type-Options: nosniff

Cache-Control

To prevent or limit the the forensic traces left on the device used by whistleblowers and in the devices involved in the communication the platform, as by section 3. Storing Responses in Caches of RFC 7234 the platform uses the Cache-control HTTP header with the configuration no-store to instruct clients and possible network proxies to disable any sort of data cache.

Cache-Control: no-store

Crawlers Policy

For security reasons the backend instructs crawlers to avoid any caching and indexing of the application and uses the Robots.txt file to enable crawling only of the home page; indexing of the home page is in fact considered best practice in order to be able to widespread the information about the existance of the platform and ease access to possible whistleblowers.

The configuration implemented is the following:

User-agent: *
Allow: /$
Disallow: *

As well the platform instruct crawlers to not keep any cache by injecting the following HTTP header:

X-Robots-Tag: noarchive

For high sensitive projects where the platform is intended to remain hidden and communicated to possible whistleblowers directly the platform could be as well configured to disable indexing completely.

The following is the HTTP header injected in this case:

X-Robots-Tag: noindex

Anchor Tags and External URLs

The client opens external urls on a new tab independent from the application context by setting rel='noreferrer' and target='_blank'` on every anchor tag.

<a href="url" rel="noreferrer" target="_blank">link title</a>

Input Validation

On the Server

The system adopts a whitelist based input validation approach. Each client request is checked against a set of regular expressions and only requests matching the expression are then processed.

As well a set of rules are applied to each request type to limit possible attacks. For example any request is limited to a payload of 1MB.

On the Client

The client implement strict validation of the rendered content by using the angular component ngSanitize.$sanitize

Form Autocomplete OFF

Form implemented by the platform make use of the HTML5 form attribute in order to instruct the browser to do not keep caching of the user data in order to predict and autocomplete forms on subsequent submissions.

This is achieved by setting autocomplete=”off” on the relevant forms or attributes.

Network Security

Connection Anonymity

Users’s anonymity is offered by means of the implementation of the Tor technology. The application implements an Onion Service v3 and advices users to use the Tor Browser when accessing to it.

Connection Encryption

Users’ connection is always encrypted, by means of the Tor Protocol while using the Tor Browser or by means of TLS when the application is accessed via a common browser.

The use of the Tor is recommended over HTTPS for its advanced properties of resistance to selective interception and censorship that would make it difficult for a third party to selectively capture or block access to the site to specific whistleblower or company department.

The software enables as well easy setup of HTTPS offering both automatic setup via Let’sEncrypt and manual setup.

TLS Certificates are generated using using NIST Curve P-384.

The configuration enables only TLS1.2+ and is fine tuned and hardened to achieve SSLLabs grade A+.

In particular only following ciphertexts are enabled:

TLS13-AES-256-GCM-SHA384
TLS13-CHACHA20-POLY1305-SHA256
TLS13-AES-128-GCM-SHA256
ECDHE-ECDSA-AES256-GCM-SHA384
ECDHE-RSA-AES256-GCM-SHA384
ECDHE-ECDSA-CHACHA20-POLY1305
ECDHE-RSA-CHACHA20-POLY1305
ECDHE-ECDSA-AES128-GCM-SHA256
ECDHE-RSA-AES128-GCM-SHA256

Network Sandboxing

The GlobaLeaks backend integrates iptables by default and implements strict firewall rules that restrict network incoming network connection to HTTP and HTTPS connection on ports 80 and 443.

In addition the application makes it possible to anonymize outgoing connections that could be configured to be sent through Tor.

Data Encryption

Submissions data, file attachment, messages and metadata exchanged between whistleblowers and recipients is encrypted using the GlobaLeaks Encryption Protocol.

In addition to this GlobaLeaks implements many other encryption components and the following is the set of the main libraries and their main usage:

Application Sandboxing

The GlobaLeaks backend integrates AppArmor by default and implements a strict sandboxing profile enabling the application to access only the strictly required files. As well the application does run under a dedicated user and group “globaleaks” with reduced privileges.

DoS Resiliency

To avoid applicative and database denial of service, GlobaLeaks apply the following measures:

  • It tries to limit the possibility of automating any operation by requiring human interaction (e.g. with the implementation of proof of work)

  • It is written to limit the possibility of triggering CPU intensive routines by an external user (e.g. by implementing limits on queries and jobs execution time)

  • It implements monitoring of each activity trying to implement detection of attacks and implement proactively security measures to prevent DoS (e.g. implementing slowdown on fast-operations)

Other Measures

Browser History and Forensic Traces

The whole application is designed keeping in mind to try to avoid or reduce the forensic traces left by whistleblowers on their devices while filing their reports.

When the accessed via the Tor Browser, the browser guarantees that no persistent traces are left on the device of the user.

In order to prevent or limit the forensic traces left in the browser history of the users accessing the platform via a common browser, the application avoids to change URI during whistleblower navigation. This has the effect to prevent the browser to log the activities performed by the user and offers high plausible deniability protection making the whistleblower appear as a simple visitor of the homepage and avoiding an actual evidence of any submission.

Secure File Download

Malware Isolation

Any attachment file uploaded by anonymous whistleblowers could possibly contain malware. To support users to download a copy of the report and protect them while transfering files to an air-gapped machine in ordr to access them, the export functionality of reports provides files wrapped in a safe Zip archive.

Download Encryption

Any attachment file uploaded by anonymous whistlblowers is protected and stored on the application server with an automated Requirements. To protect at rest files downloaded by recipients on their own computers, the platforms offers users the possibility to load their own PGP key; whenever users load their own PGP key, their downloads will be served encrypted.

Encryption of Temporary Files

Files being uploaded and temporarily stored on the disk during the upload process are encrypted with a temporary, symmetric AES-key in order to avoid writing any part of an unencrypted file’s data chunk to disk. The encryption is done in “streaming” by using AES 128bit in CTR mode. The key files are stored in memory and are unique for each file being uploaded.

Secure File Delete

Every file deleted by the application if overwritten before releasing the file space on the disk.

The overwrite routine is performed by a periodic scheduler and acts as following:

  • A first overwrite writes 0 on the whole file;

  • A second overwrite writes 1 on the whole file;

  • A third overwrite writes random bytes on the whole file.

Secure Deletion of Database Entries

The platform enables the SQLite capability for secure deletion that automatically makes the database overwrite the data upon each delete query:

PRAGMA secure_delete = ON
PRAGMA auto_vacuum = FULL

Exception Logging and Redaction

In order to quickly diagnose potential problems in the software when exceptions in clients are generated, they are automatically reported to the backend. The backend backend temporarily caches these exceptions and sends them to the backend administrator via email.

In order to prevent inadvertent information leaks the logs are run through filters that redact email addresses and uuids.

Entropy Sources

The main source of entropy for the platform is /dev/urandom.

In order to increase the entropy available on the system the system integrates the usage of the Haveged daemon.

UUIDv4 Randomness

Resources in the system like submissions and files are identified by a UUIDv4 in order to not be guessable by an external user and limit possible attacks.

TLS for SMTP Notification

All of the notifications are sent through SMTP over TLS encrypted channel by using SMTP/TLS or SMTPS, depending on the configuration.