Cybersecurity researchers have found malicious Google Chrome extensions that include capabilities to hijack affiliate hyperlinks, steal knowledge, and acquire OpenAI ChatGPT authentication tokens.
One of many extensions in query is Amazon Advertisements Blocker (ID: pnpchphmplpdimbllknjoiopmfphellj), which claims to be a device to browse Amazon with none sponsored content material. It was uploaded to the Chrome Net Retailer by a writer named “10Xprofit” on January 19, 2026.
“The extension does block adverts as marketed, however its main operate is hidden: it mechanically injects the developer’s affiliate tag (10xprofit-20) into each Amazon product hyperlink and replaces present affiliate codes from content material creators,” Socket safety researcher Kush Pandya mentioned.
Additional evaluation has decided that Amazon Advertisements Blocker is an element of a bigger cluster of 29 browser add-ons that focus on a number of e-commerce platforms like AliExpress, Amazon, Greatest Purchase, Shein, Shopify, and Walmart. The entire listing is as follows –
- AliExpress Bill Generator (FREE) – AliInvoice™️ (10+ Templates) (ID: mabbblhhnmlckjbfppkopnccllieeocp)
- AliExpress Worth Tracker – Worth Historical past & Alerts (ID: loiofaagnefbonjdjklhacdhfkolcfgi)
- AliExpress Fast Foreign money & Worth Converter (ID: mcaglpclodnaiimhicpjemhcinjfnjce)
- AliExpress Offers Countdown – Flash Sale Timer (ID: jmlgkeaofknfmnbpmlmadnfnfajdlehn)
- 10Xprofit – Amazon Vendor Instruments (FBA & FBM) (ID: ahlnchhkedmjbdocaamkbmhppnligmoh)
- Amazon Advertisements Blocker (ID: pnpchphmplpdimbllknjoiopmfphellj)
- Amazon ASIN Lookup 10xprofit (ID: ljcgnobemekghgobhlplpehijemdgcgo)
- Amazon Search Suggestion (ID: dnmfcojgjchpjcmjgpgonmhccibjopnb)
- Amazon Product Scraper 10xprofit (ID: mnacfoefejolpobogooghoclppjcgfcm)
- Amazon Fast Model Search (ID: nigamacoibifjohkmepefofohfedblgg)
- Amazon Inventory Checker 999 (ID: johobikccpnmifjjpephegmfpipfbfme)
- Amazon Worth Historical past Saver (ID: kppfbknppimnoociaomjcdgkebdmenkh)
- Amazon ASIN Copy (ID: aohfjaadlbiifnnajpobdhokecjokhab)
- Amazon Key phrase Cloud Generator (ID: gfdbbmngalhmegpkejhidhgdpmehlmnd)
- Amazon Picture Downloader (ID: cpcojeeblggnjjgnpiicndnahfhjdobd)
- Amazon Damaging Overview Hider (ID: hkkkipfcdagiocekjdhobgmlkhejjfoj)
- Amazon Itemizing Rating Checker (ID: jaojpdijbaolkhkifpgbjnhfbmckoojh)
- Amazon Key phrase Density Searcher (ID: ekomkpgkmieaaekmaldmaljljahehkoi)
- Amazon Sticky Notes (ID: hkhmodcdjhcidbcncgmnknjppphcpgmh)
- Amazon End result Numbering (ID: nipfdfkjnidadibpbflijepbllfkokac)
- Amazon Revenue Calculator Lite (ID: behckapcoohededfbgjgkgefgkpodeho)
- Amazon Weight Converter (ID: dfnannaibdndmkienngjahldiofjbkmj)
- Amazon BSR Quick View (ID: nhilffccdbcjcnoopblecppbhalagpaf)
- Amazon Character Rely & Vendor Instruments (ID: goikoilmhcgfidolicnbgggdpckdcoam)
- Amazon World Worth Checker (ID: mjcgfimemamogfmekphcfdehfkkbmldn)
- BestBuy Search By Picture (ID: nppjmiadmakeigiagilkfffplihgjlec)
- SHEIN Search By Picture (ID: mpgaodghdhmeljgogbeagpbhgdbfofgb)
- Shopify Search By Picture (ID: gjlbbcimkbncedhofeknicfkhgaocohl)
- Walmart Search By Picture (ID: mcaihdkeijgfhnlfcdehniplmaapadgb)
Whereas “Amazon Advertisements Blocker” provides the marketed performance, it additionally embeds malicious code that scans all Amazon product URL patterns for any affiliate tag with out requiring any person interplay, and replaces it with “10xprofit-20” (or “_c3pFXV63” for AliExpress). In instances the place there are not any tags, the attacker’s tag is appended to every URL.
Socket additionally famous that the extension itemizing web page on the Chrome Net Retailer makes deceptive disclosures, claiming that the builders earn a “small fee” each time a person makes use of a coupon code to make a purchase order.
Affiliate hyperlinks are extensively used throughout social media and web sites. They discuss with URLs containing a selected ID that permits monitoring of visitors and gross sales to a specific marketer. When a person clicks this hyperlink to purchase the product, the affiliate earns a reduce of the sale.
As a result of extensions trying to find present tags and changing them, social media content material creators who share Amazon product hyperlinks with their very own affiliate tags lose commissions when customers who’ve put in the add-on click on these hyperlinks.
This quantities to a violation of Chrome Net Retailer insurance policies, as they require extensions utilizing affiliate hyperlinks to precisely disclose how this system works, require person motion earlier than every injection, and by no means substitute present affiliate codes.
“The disclosure describes a coupon/deal extension with user-triggered reveals. The precise product is an advert blocker with computerized hyperlink modification,” Pandya defined. “This mismatch between disclosure and implementation creates false consent.”
“The extension additionally violates the Single Function coverage by combining two unrelated features (advert blocking and affiliate injection) that ought to be separate extensions.”
The recognized extensions have additionally been discovered to scrape product knowledge and exfiltrate it to “app.10xprofit[.]io,” with these specializing in AliExpress serving bogus “LIMITED TIME DEAL” countdown timers on product pages to create a false sense of urgency and rush them into making purchases in order to earn commissions on affiliate hyperlinks.
“Extensions that mix unrelated performance (advert blocking, value comparability, coupon discovering) with affiliate injection ought to be handled as high-risk, notably these with disclosures that do not match the precise code conduct,” Socket mentioned.
The disclosure comes as Broadcom-owned Symantec flagged 4 completely different extensions which have a mixed person base exceeding 100,000 customers and are designed to steal knowledge –
- Good Tab (ID: glckmpfajbjppappjlnhhlofhdhlcgaj), which grants full clipboard permissions to an exterior area (“api.office123456[.]com”) to allow distant clipboard-read and clipboard-write permissions
- Youngsters Safety (ID: giecgobdmgdamgffeoankaipjkdjbfep), which implements performance to reap cookies, inject adverts, and execute arbitrary JavaScript by contacting a distant server
- DPS Websafe (ID: bjoddpbfndnpeohkmpbjfhcppkhgobcg), which modifications the default search to 1 underneath their management to seize search phrases entered by customers and probably route them to malicious web sites
- Inventory Informer (ID: beifiidafjobphnbhbbgmgnndjolfcho), which is vulnerable to a years-old cross-site (XSS) vulnerability within the Stockdio Historic Chart WordPress plugin (CVE-2020-28707, CVSS rating: 6.1) that might permit a distant attacker to execute JavaScript code
“Whereas browser extensions can present a variety of helpful instruments to assist us obtain extra on-line, a lot care must be taken when selecting to put in them, even when putting in from trusted sources,” researchers Yuanjing Guo and Tommy Dong mentioned.
Rounding off the listing of malicious extensions is one other community of 16 add-ons (15 on the Chrome Net Retailer and one on the Microsoft Edge Add-ons market) which can be designed to intercept and steal ChatGPT authentication tokens by injecting a content material script into chatgpt[.]com. Cumulatively, the extensions have been downloaded about 900 instances, in line with LayerX.
The extensions are assessed to be a part of a coordinated marketing campaign as a result of overlaps in supply code, icons, branding, and descriptions –
- ChatGPT folder, voice obtain, immediate supervisor, free instruments – ChatGPT Mods (ID: lmiigijnefpkjcenfbinhdpafehaddag)
- ChatGPT voice obtain, TTS obtain – ChatGPT Mods (ID: obdobankihdfckkbfnoglefmdgmblcld)
- ChatGPT pin chat, bookmark – ChatGPT Mods (ID: kefnabicobeigajdngijnnjmljehknjl)
- ChatGPT message navigator, historical past scroller – ChatGPT Mods (ID: ifjimhnbnbniiiaihphlclkpfikcdkab)
- ChatGPT mannequin swap, save superior mannequin makes use of – ChatGPT Mods (ID: pfgbcfaiglkcoclichlojeaklcfboieh)
- ChatGPT export, Markdown, JSON, pictures – ChatGPT Mods (ID: hljdedgemmmkdalbnmnpoimdedckdkhm)
- ChatGPT Timestamp Show – ChatGPT Mods (ID: afjenpabhpfodjpncbiiahbknnghabdc)
- ChatGPT bulk delete, Chat supervisor – ChatGPT Mods (ID: gbcgjnbccjojicobfimcnfjddhpphaod)
- ChatGPT search historical past, find particular messages – ChatGPT Mods (ID: ipjgfhcjeckaibnohigmbcaonfcjepmb)
- ChatGPT immediate optimization – ChatGPT Mods (ID: mmjmcfaejolfbenlplfoihnobnggljij)
- Collapsed message – ChatGPT Mods (ID: lechagcebaneoafonkbfkljmbmaaoaec)
- Multi-Profile Administration & Switching – ChatGPT Mods (ID: nhnfaiiobkpbenbbiblmgncgokeknnno)
- Search with ChatGPT – ChatGPT Mods (ID: hpcejjllhbalkcmdikecfngkepppoknd)
- ChatGPT Token counter – ChatGPT Mods (ID: hfdpdgblphooommgcjdnnmhpglleaafj)
- ChatGPT Immediate Supervisor, Folder, Library, Auto Ship – ChatGPT Mods (ID: ioaeacncbhpmlkediaagefiegegknglc)
- ChatGPT Mods – Folder Voice Obtain & Extra Free Instruments (ID: jhohjhmbiakpgedidneeloaoloadlbdj)
With synthetic intelligence (AI)-related extensions changing into more and more frequent in enterprise workflows, the event highlights an rising assault floor the place menace actors weaponize the belief related to well-liked AI manufacturers to deceive customers into putting in them.
As a result of such instruments typically require elevated execution context inside the browser and have entry to delicate knowledge, seemingly innocent extensions can turn out to be a profitable assault vector, allowing adversaries to acquire persistent entry with out the necessity for exploiting safety flaws or resorting to different strategies which will set off safety alarms.
“Possession of such tokens gives account-level entry equal to that of the person, together with entry to dialog historical past and metadata,” safety researcher Natalie Zargarov mentioned. “Because of this, attackers can replicate the customers’ entry credentials to ChatGPT and impersonate them, permitting them to entry the entire person’s ChatGPT conversations, knowledge, or code.”
Browsers Turn out to be a Profitable Assault Vector
The findings additionally coincide with the emergence of a brand new malware-as-a-service toolkit referred to as Stanley that is being peddled on a Russian cybercrime discussion board for between $2,000 and $6,000, and permits crooks to generate malicious Chrome browser extensions that can be utilized to serve phishing pages inside an HTML iframe factor whereas nonetheless exhibiting the reputable URL within the deal with bar.
Prospects of the device acquire entry to a C2 panel for managing victims, configuring spoofed redirects, and sending pretend browser notifications. Those that are keen to spend $6,000 get a assure that any extension they create utilizing the package will move Google’s vetting course of for the Chrome Net Retailer.
These extensions take the type of innocuous note-taking utilities to fly underneath the radar. However their malicious conduct is activated when the person navigates to an internet site of curiosity to the attacker, corresponding to a financial institution, at which level a full-screen iframe containing the phishing web page is overlaid, whereas leaving the browser’s URL bar intact. This visible deception creates a defensive blind spot that may dupe even vigilant customers into getting into their credentials or delicate data on the web page.
As of January 27, 2025, the service seems to have vanished – seemingly prompted by the general public disclosure – but it surely’s very a lot potential that it may possibly resurface underneath a distinct title sooner or later.
“Stanley gives a turnkey website-spoofing operation disguised as a Chrome extension, with its premium tier promising assured publication on the Chrome Net Retailer,” Varonis researcher Daniel Kelley famous earlier this week. “BYOD insurance policies, SaaS-first environments, and distant work have made the browser the brand new endpoint. Attackers have seen. Malicious browser extensions at the moment are a main assault vector.”
