Ecco un po’ di foto dalle nostre Cellule Coscioni in giro per l’Italia: soggetto di questo mese, il ciclo di eventi “Una dose di realtà” su salute mentale e psichedelici, che si sono svolti in Sardegna. Clicca sull’icona della macchina fotografica per scorrere le foto!

L'articolo La Gallery di Agosto proviene da Associazione Luca Coscioni.

When you visit almost any website, you’ll see a pop-up asking you to accept, decline, or customize the cookies it collects. Sometimes, it just tells you that cookies are in use by default. We randomly checked 647 websites, and 563 of them displayed cookie notifications. Most of the time, users don’t even pause to think about what’s really behind the banner asking them to accept or decline cookies.

We owe cookie warnings to the adoption of new laws and regulations, such as GDPR, that govern the collection of user information and protection of personal data. By adjusting your cookie settings, you can minimize the amount of information collected about your online activity. For example, you can decline to collect and store third-party cookies. These often aren’t necessary for a website to function and are mainly used for marketing and analytics. This article explains what cookies are, the different types, how they work, and why websites need to warn you about them. We’ll also dive into sensitive cookies that hold the Session ID, the types of attacks that target them, and ways for both developers and users to protect themselves.

What are browser cookies?

Cookies are text files with bits of data that a web server sends to your browser when you visit a website. The browser saves this data on your device and sends it back to the server with every future request you make to that site. This is how the website identifies you and makes your experience smoother.

Let’s take a closer look at what kind of data can end up in a cookie.

First, there’s information about your actions on the site and session parameters: clicks, pages you’ve visited, how long you were on the site, your language, region, items you’ve added to your shopping cart, profile settings (like a theme), and more. This also includes data about your device: the model, operating system, and browser type.

Your sign-in credentials and security tokens are also collected to identify you and make it easier for you to sign in. Although it’s not recommended to store this kind of information in cookies, it can happen, for example, when you check the “Remember me” box. Security tokens can become vulnerable if they are placed in cookies that are accessible to JS scripts.

Another important type of information stored in cookies that can be dangerous if it falls into the wrong hands is the Session ID: a unique code assigned to you when you visit a website. This is the main target of session hijacking attacks because it allows an attacker to impersonate the user. We’ll talk more about this type of attack later. It’s worth noting that a Session ID can be stored in cookies, or it can even be written directly into the URL of the page if the user has disabled cookies.

Example of a Session ID as displayed in the Firefox browser’s developer panel

Example of a Session ID as seen in a URL address: example.org/?account.php?osCsid=dawnodpasb<...>abdisoa.

Besides the information mentioned above, cookies can also hold some of your primary personal data, such as your phone number, address, or even bank card details. They can also inadvertently store confidential company information that you’ve entered on a website, including client details, project information, and internal documents.

Many of these data types are considered sensitive. This means if they are exposed to the wrong people, they could harm you or your organization. While things like your device type and what pages you visited aren’t typically considered confidential, they still create a detailed profile of you. This information could be used by attackers for phishing scams or even blackmail.

Main types of cookies

Cookies by storage time

Cookies are generally classified based on how long they are stored. They come in two main varieties: temporary and persistent.

Temporary, or session cookies, are used during a visit to a website and deleted as soon as you leave. They save you from having to sign in every time you navigate to a new page on the same site or to re-select your language and region settings. During a single session, these values are stored in a cookie because they ensure uninterrupted access to your account and proper functioning of the site’s features for registered users. Additionally, temporary cookies include things like entries in order forms and pages you visited. This information can end up in persistent cookies if you select options like “Remember my choice” or “Save settings”. It’s important to note that session cookies won’t get deleted if you have your browser set to automatically restore your previous session (load previously opened tabs). In this case, the system considers all your activity on that site as one session.

Persistent cookies, unlike temporary ones, stick around even after you leave the site. The website owner sets an expiration date for them, typically up to a year. You can, however, delete them at any time by clearing your browser’s cookies. These cookies are often used to store sign-in credentials, phone numbers, addresses, or payment details. They’re also used for advertising to determine your preferences. Sensitive persistent cookies often have a special attribute HttpOnly. This prevents your browser from accessing their contents, so the data is sent directly to the server every time you visit the site.

Notably, depending on your actions on the website, credentials may be stored in either temporary or persistent cookies. For example, when you simply navigate a site, your username and password might be stored in session cookies. But if you check the “Remember me” box, those same details will be saved in persistent cookies instead.

Cookies by source

Based on the source, cookies are either first-party or third-party. The former are created and stored by the website, and the latter, by other websites. Let’s take a closer look at these cookie types.

First-party cookies are generally used to make the site function properly and to identify you as a user. However, they can also perform an analytics or marketing function. When this is the case, they are often considered optional – more on this later – unless their purpose is to track your behavior during a specific session.

Third-party cookies are created by websites that the one you’re visiting is talking to. The most common use for these is advertising banners. For example, a company that places a banner ad on the site can use a third-party cookie to track your behavior: how many times you click on the ad and so on. These cookies are also used by analytics services like Google Analytics or Yandex Metrica.

Social media cookies are another type of cookies that fits into this category. These are set by widgets and buttons, such as “Share” or “Like”. They handle any interactions with social media platforms, so they might store your sign-in credentials and user settings to make those interactions faster.

Cookies by importance

Another way to categorize cookies is by dividing them into required and optional.

Required or essential cookies are necessary for the website’s basic functions or to provide the service you’ve specifically asked for. This includes temporary cookies that track your activity during a single visit. It also includes security cookies, such as identification cookies, which the website uses to recognize you and spot any fraudulent activity. Notably, cookies that store your consent to save cookies may also be considered essential if determined by the website owner, since they are necessary to ensure the resource complies with your chosen privacy settings.

The need to use essential cookies is primarily relevant for websites that have a complex structure and a variety of widgets. Think of an e-commerce site that needs a shopping cart and a payment system, or a photo app that has to save images to your device.

A key piece of data stored in required cookies is the above-mentioned Session ID, which helps the site identify you. If you don’t allow this ID to be saved in a cookie, some websites will put it directly in the page’s URL instead. This is a much riskier practice because URLs aren’t encrypted. They’re also visible to analytics services, tracking tools, and even other users on the same network as you, which makes them vulnerable to cross-site scripting (XSS) attacks. This is a major reason why many sites won’t let you disable required cookies for your own security.

Example of required cookies on the Osano CMP website

Optional cookies are the ones that track your online behavior for marketing, analytics, and performance. This category includes third-party cookies created by social media platforms, as well as performance cookies that help the website run faster and balance the load across servers. For instance, these cookies can track broken links to improve a website’s overall speed and reliability.

Essentially, most optional cookies are third-party cookies that aren’t critical for the site to function. However, the category can also include some first-party cookies for things like site analytics or collecting information about your preferences to show you personalized content.

While these cookies generally don’t store your personal information in readable form, the data they collect can still be used by analytics tools to build a detailed profile of you with enough identifying information. For example, by analyzing which sites you visit, companies can make educated guesses about your age, health, location, and much more.

A major concern is that optional cookies can sometimes capture sensitive information from autofill forms, such as your name, home address, or even bank card details. This is exactly why many websites now give you the choice to accept or decline the collection of this data.

Special types of cookies

Let’s also highlight special subtypes of cookies managed with the help of two similar technologies that enable non-standard storage and retrieval methods.

A supercookie is a tracking technology that embeds cookies into website headers and stores them in non-standard locations, such as HTML5 local storage, browser plugin storage, or browser cache. Because they’re not in the usual spot, simply clearing your browser’s history and cookies won’t get rid of them.

Supercookies are used for personalizing ads and collecting analytical data about the user (for example, by internet service providers). From a privacy standpoint, supercookies are a major concern. They’re a persistent and hard-to-control tracking mechanism that can monitor your activity without your consent, which makes it tough to opt out.

Another unusual tracking method is Evercookie, a type of zombie cookie. Evercookies can be recovered with JavaScript even after being deleted. The recovery process relies on the unique user identifier (if available), as well as traces of cookies stored across all possible browser storage locations.

How cookie use is regulated

The collection and management of cookies are governed by different laws around the world. Let’s review the key standards from global practices.

1. General Data Protection Regulation (GDPR) and ePrivacy Directive (Cookie Law) in the European Union.
   Under EU law, essential cookies don’t require user consent. This has created a loophole for some websites. You might click “Reject All”, but that button might only refuse non-essential cookies, allowing others to still be collected.
2. Lei Geral de Proteção de Dados Pessoais (LGPD) in Brazil.
   This law regulates the collection, processing, and storage of user data within Brazil. It is largely inspired by the principles of GDPR and, similarly, requires free, unequivocal, and clear consent from users for the use of their personal data. However, LGPD classifies a broader range of information as personal data, including biometric and genetic data. It is important to note that compliance with GDPR does not automatically mean compliance with LGPD, and vice versa.
3. California Consumer Privacy Act (CCPA) in the United States.
   The CCPA considers cookies a form of personal information. This means their collection and storage must follow certain rules. For example, any California resident has the right to stop cross-site cookie tracking to prevent their personal data from being sold. Service providers are required to give users choices about what data is collected and how it’s used.
4. The UK’s Privacy and Electronic Communications Regulations (PECR, or EC Directive) are similar to the Cookie Law.
   PECR states that websites and apps can only save information on a user’s device in two situations: when it’s absolutely necessary for the site to work or provide a service, or when the user has given their explicit consent to this.
5. Federal Law No. 152-FZ “On Personal Data” in Russia.
   The law broadly defines personal data as any information that directly or indirectly relates to an individual. Since cookies can fall under this definition, they can be regulated by this law. This means websites must get explicit consent from users to process their data.

In Russia, website owners must inform users about the use of technical cookies, but they don’t need to get consent to collect this information. For all other types of cookies, user consent is required. Often, the user gives this consent automatically when they first visit the site, as it’s stated in the default cookie warning.

Some sites use a banner or a pop-up window to ask for consent, and some even let users choose exactly which cookies they’re willing to store on their device.

Beyond these laws, website owners create their own rules for using first-party cookies. Similarly, third-party cookies are managed by the owners of third-party services, such as Google Analytics. These parties decide what kind of information goes into the cookies and how it’s formatted. They also determine the cookies’ lifespan and security settings. To understand why these settings are so important, let’s look at a few ways malicious actors can attack one of the most critical types of cookies: those that contain a Session ID.

Session hijacking methods

As discussed above, cookies containing a Session ID are extremely sensitive. They are a prime target for cybercriminals. In real-world attacks, different methods for stealing a Session ID have been documented. This is a practice known as session hijacking. Below, we’ll look at a few types of session hijacking.

Session sniffing

One method for stealing cookies with a Session ID is session sniffing, which involves intercepting traffic between the user and the website. This threat is a concern for websites that use the open HTTP protocol instead of HTTPS, which encrypts traffic. With HTTP, cookies are transmitted in plain text within the headers of HTTP requests, which makes them vulnerable to interception.

Attacks targeting unencrypted HTTP traffic mostly happen on public Wi-Fi networks, especially those without a password and strong security protocols like WPA2 or WPA3. These protocols use AES encryption to protect traffic on Wi-Fi networks, with WPA3 currently being the most secure version. While WPA2/WPA3 protection limits the ability to intercept HTTP traffic, only implementing HTTPS can truly protect against session sniffing.

This method of stealing Session ID cookies is fairly rare today, as most websites now use HTTPS encryption. The popularity of this type of attack, however, was a major reason for the mass shift to using HTTPS for all connections during a user’s session, known as HTTPS everywhere.

Cross-site scripting (XSS)

Cross-site scripting (XSS) exploits vulnerabilities in a website’s code to inject a malicious script, often written in JavaScript, onto its webpages. This script then runs whenever a victim visits the site. Here’s how an XSS attack works: an attacker finds a vulnerability in the source code of the target website that allows them to inject a malicious script. For example, the script might be hidden in a URL parameter or a comment on the page. When the user opens the infected page, the script executes in their browser and gains access to the site’s data, including the cookies that contain the Session ID.

Session fixation

In a session fixation attack, the attacker tricks your browser into using a pre-determined Session ID. Thus, the attacker prepares the ground for intercepting session data after the victim visits the website and performs authentication.

Here’s how it goes down. The attacker visits a website and gets a valid, but unauthenticated, Session ID from the server. They then trick you into using that specific Session ID. A common way to do this is by sending you a link with the Session ID already embedded in the URL, like this: http://example.com/?SESSIONID=ATTACKER_ID. When you click the link and sign in, the website links the attacker’s Session ID to your authenticated session. The attacker can then use the hijacked Session ID to take over your account.

Modern, well-configured websites are much less vulnerable to session fixation than XSS-like attacks because most current web frameworks automatically change the user’s Session ID after they sign in. However, the very existence of this Session ID exploitation attack highlights how crucial it is for websites to securely manage the entire lifecycle of the user session, especially at the moment of sign-in.

Cross-site request forgery (CSRF)

Unlike session fixation or sniffing attacks, cross-site request forgery (CSRF or XSRF) leverages the website’s trust in your browser. The attacker forces your browser, without your knowledge, to perform an unwanted action on a website where you’re signed in – like changing your password or deleting data.

For this type of attack, the attacker creates a malicious webpage or an email message with a harmful link, piece of HTML code, or script. This code contains a request to a vulnerable website. You open the page or email message, and your browser automatically sends the hidden request to the target site. The request includes the malicious action and all the necessary (for example, temporary) cookies for that site. Because the website sees the valid cookies, it treats the request as a legitimate one and executes it.

Variants of the man-in-the-middle (MitM) attack

A man-in-the-middle (MitM) attack is when a cybercriminal not only snoops on but also redirects all the victim’s traffic through their own systems, thus gaining the ability to both read and alter the data being transmitted. Examples of these attacks include DNS spoofing or the creation of fake Wi-Fi hotspots that look legitimate. In an MitM attack, the attacker becomes the middleman between you and the website, which gives them the ability to intercept data, such as cookies containing the Session ID.

Websites using the older HTTP protocol are especially vulnerable to MitM attacks. However, sites using the more secure HTTPS protocol are not entirely safe either. Malicious actors can try to trick your browser with a fake SSL/TLS certificate. Your browser is designed to warn you about suspicious invalid certificates, but if you ignore that warning, the attacker can decrypt your traffic. Cybercriminals can also use a technique called SSL stripping to force your connection to switch from HTTPS to HTTP.

Predictable Session IDs

Cybercriminals don’t always have to steal your Session ID – sometimes they can just guess it. They can figure out your Session ID if it’s created according to a predictable pattern with weak, non-cryptographic characters. For example, a Session ID may contain your IP address or consecutive numbers, and a weak algorithm that uses easily predictable random sequences may be used to generate it.

To carry out this type of attack, the malicious actor will collect a sufficient number of Session ID examples. They analyze the pattern to figure out the algorithm used to create the IDs, then apply that knowledge to predicting your current or next Session ID.

Cookie tossing

This attack method exploits the browser’s handling of cookies set by subdomains of a single domain. If a malicious actor takes control of a subdomain, they can try to manipulate higher-level cookies, in particular the Session ID. For example, if a cookie is set for sub.domain.com with the Domain attribute set to .domain.com, that cookie will also be valid for the entire domain.

This lets the attacker “toss” their own malicious cookies with the same names as the main domain’s cookies, such as Session_id. When your browser sends a request to the main server, it includes all the relevant cookies it has. The server might mistakenly process the hacker’s Session ID, giving them access to your user session. This can work even if you never visited the compromised subdomain yourself. In some cases, sending invalid cookies can also cause errors on the server.

How to protect yourself and your users

The primary responsibility for cookie security rests with website developers. Modern ready-made web frameworks generally provide built-in defenses, but every developer should understand the specifics of cookie configuration and the risks of a careless approach. To counter the threats we’ve discussed, here are some key recommendations.

Recommendations for web developers

All traffic between the client and server must be encrypted at the network connection and data exchange level. We strongly recommend using HTTPS and enforcing automatic redirect from HTTP to HTTPS. For an extra layer of protection, developers should use the HTTP Strict Transport Security (HSTS) header, which forces the browser to always use HTTPS. This makes it much harder, and sometimes impossible, for attackers to slip into your traffic to perform session sniffing, MitM, or cookie tossing attacks.

It must be mentioned that the use of HTTPS is insufficient protection against XSS attacks. HTTPS encrypts data during transmission, while an XSS script executes directly in the user’s browser within the HTTPS session. So, it’s up to the website owner to implement protection against XSS attacks. To stop malicious scripts from getting in, developers need to follow secure coding practices:

• Validate and sanitize user input data.
• Implement mandatory data encoding (escaping) when rendering content on the page – this way, the browser will not interpret malicious code as part of the page and will not execute it.
• Use the HttpOnly flag to protect cookie files from being accessed by the browser.
• Use the Content Security Policy (CSP) standard to control code sources. It allows monitoring which scripts and other content sources are permitted to execute and load on the website.

For attacks like session fixation, a key defense is to force the server to generate a new Session ID right after the user successfully signs in. The website developer must invalidate the old, potentially compromised Session ID and create a new one that the attacker doesn’t know.

An extra layer of protection involves checking cookie attributes. To ensure protection, it is necessary to check for the presence of specific flags (and set them if they are missing): Secure and HttpOnly. The Secure flag ensures that cookies are transmitted over an HTTPS connection, while HttpOnly prevents access to them from the browser, for example through scripts, helping protect sensitive data from malicious code. Having these attributes can help protect against session sniffing, MitM, cookie tossing, and XSS.

Pay attention to another security attribute, SameSite, which can restrict cookie transmission. Set it to Lax or Strict for all cookies to ensure they are sent only to trusted web addresses during cross-site requests and to protect against CSRF attacks. Another common strategy against CSRF attacks is to use a unique, randomly generated CSRF token for each user session. This token is sent to the user’s browser and must be included in every HTTP request that performs an action on your site. The site then checks to make sure the token is present and correct. If it’s missing or doesn’t match the expected value, the request is rejected as a potential threat. This is important because if the Session ID is compromised, the attacker may attempt to replace the CSRF token.

To protect against an attack where a cybercriminal tries to guess the user’s Session ID, you need to make sure these IDs are truly random and impossible to predict. We recommend using a cryptographically secure random number generator that utilizes powerful algorithms to create hard-to-predict IDs. Additional protection for the Session ID can be ensured by forcing its regeneration after the user authenticates on the web resource.

The most effective way to prevent a cookie tossing attack is to use cookies with the __Host- prefix. These cookies can only be set on the same domain that the request originates from and cannot have a Domain attribute specified. This guarantees that a cookie set by the main domain can’t be overwritten by a subdomain.

Finally, it’s crucial to perform regular security checks on all your subdomains. This includes monitoring for inactive or outdated DNS records that could be hijacked by an attacker. We also recommend ensuring that any user-generated content is securely isolated on its own subdomain. User-generated data must be stored and managed in a way that prevents it from compromising the security of the main domain.

As mentioned above, if cookies are disabled, the Session ID can sometimes get exposed in the website URL. To prevent this, website developers must embed this ID into essential cookies that cannot be declined.

Many modern web development frameworks have built-in security features that can stop most of the attack types described above. These features make managing cookies much safer and easier for developers. Some of the best practices include regular rotation of the Session ID after the user signs in, use of the Secure and HttpOnly flags, limiting the session lifetime, binding it to the client’s IP address, User-Agent string, and other parameters, as well as generating unique CSRF tokens.

There are other ways to store user data that are both more secure and better for performance than cookies.

Depending on the website’s needs, developers can use different tools, like the Web Storage API (which includes localStorage and sessionStorage), IndexedDB, and other options. When using an API, data isn’t sent to the server with every single request, which saves resources and makes the website perform better.

Another exciting alternative is the server-side approach. With this method, only the Session ID is stored on the client side, while all the other data stays on the server. This is even more secure than storing data with the help of APIs because private information is never exposed on the client side.

Tips for users

Staying vigilant and attentive is a big part of protecting yourself from cookie hijacking and other malicious manipulations.

Always make sure the website you are visiting is using HTTPS. You can check this by looking at the beginning of the website address in the browser address bar. Some browsers let the user view additional website security details. For example, in Google Chrome, you can click the icon right before the address.

This will show you if the “Connection is secure” and the “Certificate is valid”. If these details are missing or data is being sent over HTTP, we recommend maximum caution when visiting the website and, whenever possible, avoiding entering any personal information, as the site does not meet basic security standards.

When browsing the web, always pay attention to any security warnings your browser gives you, especially about suspicious or invalid certificates. Seeing one of these warnings might be a sign of an MitM attack. If you see a security warning, it’s best to stop what you’re doing and leave that website right away. Many browsers implement certificate verification and other security features, so it is important to install browser updates promptly – this replaces outdated and compromised certificates.

We also recommend regularly clearing your browser data (cookies and cache). This can help get rid of outdated or potentially compromised Session IDs.

Always use two-factor authentication wherever it’s available. This makes it much harder for a malicious actor to access your account, even if your Session ID is exposed.

When a site asks for your consent to use cookies, the safest option is to refuse all non-essential ones, but we’ll reiterate that sometimes, clicking “Reject cookies” only means declining the optional ones. If this option is unavailable, we recommend reviewing the settings to only accept the strictly necessary cookies. Some websites offer this directly in the pop-up cookie consent notification, while others provide it in advanced settings.

The universal recommendation to avoid clicking suspicious links is especially relevant in the context of preventing Session ID theft. As mentioned above, suspicious links can be used in what’s known as session fixation attacks. Carefully check the URL: if it contains parameters you do not understand, we recommend copying the link into the address bar manually and removing the parameters before loading the page. Long strings of characters in the parameters of a legitimate URL may turn out to be an attacker’s Session ID. Deleting it renders the link safe. While you’re at it, always check the domain name to make sure you’re not falling for a phishing scam.

In addition, we advise extreme caution when connecting to public Wi-Fi networks. Man-in-the-middle attacks often happen through open networks or rogue Wi-Fi hotspots. If you need to use a public network, never do it without a virtual private network (VPN), which encrypts your data and makes it nearly impossible for anyone to snoop on your activity.

securelist.com/cookies-and-ses…

Over the years Intel has introduced a number of new computer form factors that either became a hit, fizzled out, or moved on to live a more quiet life. The New Unit of Computing (NUC) decidedly became a hit with so-called Mini PCs now everywhere, while the Intel Compute Stick has been largely forgotten. In a recent video by the [Action Retro] one such Compute Stick is poked at, specifically the last model released by Intel in the form of the 2016-era STK1AW32SC, featuring a quad-core Intel Atom x5-Z8330 SoC, 2 GB of RAM and 32 GB eMMC storage.

As the name suggests, this form factor is very stick-like, with a design that makes it easy to just plug it into the HDMI port of a display, making it a snap to add a computer to any TV or such without taking up a considerable amount of space. Although Intel didn’t make more of them after this model, it could be argued that devices like the Chromecast dongle follow the same general concept, and manufacturers like MeLe are still making new PCs in this form factor today.

In the video this 2016-era Compute Stick is put through its paces, wiping the Windows 10 installation that was still on it from the last time it was used, and an installation of Haiku was attempted which unfortunately failed to see the eMMC storage. Worse was the current Ubuntu, which saw its installer simply freeze up, but MX Linux saved the day, providing a very usable Linux desktop experience including the watching of YouTube content and network streaming of Steam games.

Although dissed as ‘e-waste’ by many today, if anything this video shows that these little sticks are still very capable computers in 2025.

youtube.com/embed/G3WvOzdlpwY?…

hackaday.com/2025/09/02/rememb…

While you can still find tape being used for backup storage, it’s pretty safe to say that the humble audio cassette is about as out of date as a media format can be. Still, it has a certain retro charm we’re suckers for, particularly in the shape of a Commodore Datasette. We’re also suckers for miniaturization, so how could we not fall for [bitluni] ‘s tiny datasette replica?

Aesthetically, he’s copying the Commodore original to get those sweet nostalgia juices flowing, but to make things more interesting he’s not using compact cassette tapes. Instead, [bitluni] started with a mini cassette dictaphone, which he tore down to its essentials and rebuilt into the Commodore-shaped case.

The prototyping of this project was full of hacks — like building a resistor ladder DAC in an unpopulated part of a spare PCB from an unrelated project. The DAC is of course key to getting data onto the mini-casettes. After some playing around [bitluni] decided that encoding data with FSK (frequency-shift keying), as was done back on the C-64, was the way to go. (Almost like those old engineers knew what they were doing!) The dictaphone tape transport is inferior to the old Datasette, though, so as a cheap error-correction hack, [bitluni] needed to duplicate each byte to make sure it gets read correctly.

The mini cassettes only fit a laughable amount of data by modern standards this way (about 1 MB) but, of course that’s not the point. If you jump to 11:33 in the video embedded below, you can see the point: the shout of triumph when loading PacMan (all 8 kB of it) from tape via USB. That transfer was via serial console; eventually [bitluni] intends to turn this into the world’s least-practical mass storage device, but that wasn’t necessary for proof-of-concept. The code for what’s shown is available on GitHub.

If you have an old Datasette you want to use with a modern PC, you’d better believe that we’ve got you covered. We’ve seen other cassette-mass-storage interfaces over the years, too. It might be a dead medium, but there’s just something about “sticky tape and rust” that lives on in our imaginations.

youtube.com/embed/GQwTPH67YqY?…

Thanks to [Stephen Walters] for the tip.

hackaday.com/2025/09/01/tiny-d…

La recente conferma da parte di Zscaler riguardo a una violazione dati derivante da un attacco alla supply chain fornisce un caso studio sull’evoluzione delle minacce contro ecosistemi SaaS complessi. L’attacco, attribuito al gruppo APT UNC6395, ha sfruttato vulnerabilità a livello di gestione delle credenziali OAuth e di API trust model nelle integrazioni tra applicazioni di terze parti e piattaforme cloud.

Secondo le prime analisi, il punto d’ingresso è stato l’abuso dell’integrazione tra Salesloft Drift e Salesforce. L’attore ha esfiltrato token OAuth validi, consentendo l’accesso diretto agli endpoint Salesforce senza dover interagire con i sistemi di autenticazione tradizionali (es. MFA o session cookies).

Questo vettore sfrutta un punto debole intrinseco nel protocollo OAuth: i bearer token. Un bearer token, se sottratto, garantisce pieno accesso fino alla sua scadenza, indipendentemente dal contesto in cui viene utilizzato. una volta ottenuto il bearer token OAuth (es. tramite furto da log, memory dump, o intercettazione),

lo si può riutilizzare in un’altra sessione, da un altro dispositivo o da un’altra rete, senza dover conoscere la password o superare l’autenticazione a più fattori. In pratica, il token rubato diventa un “passaporto valido” fino alla sua scadenza.

Gli aggressori hanno quindi orchestrato enumerazioni automatizzate via script Python, con query massicce alle API Salesforce, ottenendo dataset contenenti email, numeri di telefono e altre informazioni di contatto business.

Analisi TTP

• Initial Access: sfruttamento dell’integrazione SaaS (Salesloft Drift → Salesforce).
• Credential Access: esfiltrazione di OAuth tokens mediante attacchi mirati a log o ambienti CI/CD compromessi.
• Defense Evasion: utilizzo di token validi per evitare alert su login anomali o tentativi MFA falliti.
• Collection: scraping massivo tramite script Python (indicatore di automazione e infrastruttura consolidata).
• Exfiltration: trasferimento dei dataset su infrastrutture C2, probabilmente distribuite su cloud provider legittimi per confondere il traffico.

Questo approccio dimostra un livello elevato di maturità operativa, con chiara attenzione alla persistenza stealth e al mascheramento nel rumore operativo SaaS.

Zscaler ha confermato che la compromissione è circoscritta all’ambiente Salesforce e non ai sistemi core di sicurezza; i dati esfiltrati riguardano informazioni di contatto business, senza impatto diretto sulle infrastrutture di rete o sui servizi SASE; non sono state rilevate manipolazioni di configurazioni o codice eseguibile.

Tuttavia, anche dati apparentemente “a basso impatto” possono costituire una base privilegiata per future operazioni di spear-phishing contro clienti e partner, sfruttando la fiducia nel brand Zscaler.

Considerazioni Architetturali: Debolezze del Modello SaaS

Questo incidente conferma criticità note ma spesso trascurate:

• Token OAuth come single point of failure: senza meccanismi di rotazione rapida, scoping rigoroso e binding contestuale, diventano equivalenti a credenziali statiche.
• API Exposure: le piattaforme SaaS esposte via API spesso mancano di un sistema granulare di rate-limiting e anomaly detection, facilitando scraping su larga scala.
• Third-Party Trust: ogni applicazione integrata amplia il perimetro di rischio; la security posture dell’intero ecosistema dipende dall’anello più debole.
• Visibility Gaps: i log OAuth e le audit trail delle API non sempre sono correttamente centralizzati in SIEM, riducendo la capacità di detection.

Mitigazioni Tecniche e Best Practice

• Token hardening: implementazione di PKCE (Proof Key for Code Exchange) e token binding per ridurre il rischio di replay.
• Rotazione aggressiva: token a breve durata, con refresh gestito tramite canali sicuri.
• Scope minimization: limitare i privilegi dei token al minimo necessario (principio del least privilege).
• API monitoring: anomalie nelle chiamate API (es. spike di richieste, query massive) devono triggerare alert in tempo reale.
• Zero Trust enforcement: anche per le applicazioni interne e SaaS, ogni accesso deve essere verificato dinamicamente in base al contesto.

Conclusioni

L’attacco subito da Zscaler non rappresenta solo un incidente isolato, ma un campanello d’allarme per tutto il settore. Le architetture SaaS, per loro natura interconnesse, erodono il concetto tradizionale di perimetro. In questo scenario, la resilienza dipende dalla capacità di gestire in maniera proattiva tokenization, API exposure e terze parti.

Il caso Zscaler dimostra che anche player globali della sicurezza non sono immuni da vulnerabilità di supply chain. Il futuro della sicurezza cloud richiede un cambio di paradigma: trattare ogni integrazione come un potenziale threat vector e applicare controlli di sicurezza by design a ogni layer della catena di fornitura digitale.

L'articolo Zscaler Violazione Dati: Lezione Apprese sull’Evoluzione delle Minacce SaaS proviene da il blog della sicurezza informatica.

Un attacco informatico di vasta portata ha preso di mira l’azienda di sicurezza Zscaler, la quale ha ufficialmente confermato di essere stata vittima di una violazione nella catena di approvvigionamento. Questo attacco ha portato all’esposizione dei dati di contatto dei clienti a causa di credenziali Salesforce compromesse, collegate alla piattaforma di marketing Salesloft Drift. L’evento, reso pubblico il 31 agosto 2025, rappresenta il risultato di una campagna più vasta che ha avuto come obiettivo i token OAuth di Salesloft Drift, coinvolgendo oltre 700 organizzazioni a livello globale.

La violazione è dovuta a un attacco più ampio alla supply chain su Salesloft Drift, in cui gli autori della minaccia hanno rubato token OAuth e di aggiornamento. Questi token hanno concesso l’accesso non autorizzato alle istanze dei clienti Salesforce, consentendo l’esfiltrazione di informazioni sensibili. Nel suo avviso, Zscaler ha confermato che la sua istanza Salesforce era tra quelle interessate.

“Nell’ambito di questa campagna, soggetti non autorizzati hanno ottenuto l’accesso alle credenziali Salesloft Drift dei suoi clienti, tra cui Zscaler”, ha dichiarato l’azienda. “A seguito di un’analisi dettagliata, abbiamo stabilito che queste credenziali consentivano un accesso limitato ad alcuni dati Salesforce”.

Le informazioni esposte sono le seguenti:

• nomi dei clienti
• Indirizzi email aziendali
• Titoli di lavoro
• Numeri di telefono
• Dettagli regionali/località
• Licenza del prodotto e dettagli commerciali
• Contenuto da alcuni casi di assistenza clienti

La società Zscaler ha fatto sapere che la violazione dei dati ha riguardato esclusivamente il suo sistema Salesforce, escludendo quindi qualsiasi impatto su prodotti, infrastrutture o servizi offerti dalla stessa Zscaler. Nonostante ad oggi non siano emersi casi di abuso, la compagnia ha raccomandato ai propri clienti di rimanere vigili nei confronti di possibili tentativi di phishing o di ingegneria sociale che potrebbero sfruttare le informazioni trapelate.

Zscaler ha messo in atto una serie di misure di mitigazione per contenere l’incidente: sono state revocate tutte le integrazioni di Salesloft Drift con Salesforce, i token API sono stati ruotati per prevenire abusi futuri e, per ridurre i rischi legati all’ingegneria sociale, è stata introdotta un’autenticazione avanzata dei clienti durante le chiamate di supporto. L’azienda ha inoltre confermato che le indagini sull’accaduto restano in corso per identificare con precisione l’entità della compromissione e garantire la piena sicurezza delle integrazioni.

Google Threat Intelligence ha attribuito la compromissione di Drift al gruppo UNC6395, responsabile del furto di casi di supporto Salesforce con l’obiettivo di raccogliere credenziali, chiavi di accesso AWS, token Snowflake e altri dati sensibili. Secondo i ricercatori, gli attaccanti hanno dimostrato avanzate tattiche di sicurezza operativa, come la cancellazione dei processi di query per nascondere le proprie attività, anche se i log sono rimasti disponibili per le analisi forensi. La campagna, tuttavia, non si è limitata all’integrazione Drift–Salesforce: gli hacker hanno compromesso anche Drift Email, ottenendo l’accesso a dati di CRM e marketing automation, oltre a sfruttare token OAuth rubati per infiltrarsi negli account Google Workspace e leggere le email aziendali.

L'articolo Zscaler subisce violazione dati: attacco supply chain tramite Salesloft Drift proviene da il blog della sicurezza informatica.

Narratives around age verification and restriction of access for minors are gaining traction in the EU, amid similar efforts being pursued in the UK, US and Australia. This blog analyses different EU policy files and warns that relying on age-gating risks undermining more holistic, rights-respecting and effective solutions to online harm.

The post Age verification gains traction: the EU risks failing to address the root causes of online harm appeared first on European Digital Rights (EDRi).

PRESIDIO ASSEMBLEA - RACCOLTA FIRME
AL SITO
2 SETTEMBRE ORE 18.00

#Ambiente #StopInceneritore #NoInceneritore #NoInceneritori #ZeroWaste #Rifiuti #Riciclo #EconomiaCircolare #NoAlCarbone #EnergiaPulita

Thanks to everyone who came out to oppose the Boston Police Department’s (BPD) request for City Council approval for BPD to use three new social media surveillance tools. Because of your effort, the City Council voted against this proposal. The BPD started using these tools claiming “exigent circumstances” months before asking for approval.

We know that any tool BPD uses will feed into the Boston Regional Information Center (BRIC) and Federal agencies such as ICE, CBP and the FBI. Those tools will be used to spy on and abuse people who are doing nothing wrong. We are happy that the Boston City Council agreed.

Councilors Breadon (District 9), Louijeune (At large), Mejia (At large), Pepén (District 5), Weber (District 6) and Worrell (District 4) voted to reject the report, sending it to an oversight board that will assess whether Boston Police overstepped the surveillance ordinance. Councilors Durkan (District 8), FitzGerald (District 3), Flynn (District 2), Murphy (At large) and Santana (At large) voted to accept the report and allow BPD to continue to spy on the residents of Greater Boston.

We hope Bostonians consider how these councilors voted when they cast their ballot in the September 9th preliminary and November 4th general election. For the:

• Preliminary election: It is too late to register to vote. Early voting is September 2 – 5 and mail voting is open.
• General election: The voter registration deadline is October 25, 2025. Early voting will be October 25 – 31 (Sites and times to be determined)

Polls are open 7 am – 8 pm on election days. Boston seeks poll workers. Volunteers must attend a two-hour paid training and earn stipends of $160 – $200. Bilingual speakers are strongly encouraged to apply. Sign up at their portal.

masspirates.org/blog/2025/09/0…