Loading story…
Reporting described data brokers and “privacy” browser extensions harvesting and monetizing users’ chatbot conversations with services such as ChatGPT, Gemini, Claude, and DeepSeek. The technique cited involves extensions overriding browser fetch() and XMLHttpRequest() to capture prompts and responses, storing them in a vector database and exposing them via API; while IDs may be pseudonymized (e.g., SHA-256 hashes), the captured text can include highly sensitive data (names, dates of birth, medical record numbers, diagnosis codes), creating re-identification and privacy risks.
Separately, dark-web monitoring reported multiple alleged data leak/sale claims and a malware offering: an actor claimed to have exfiltrated an Eholo healthcare-related database (including large volumes of medical notes and PII) and demanded a ransom with a threat to sell the data, and another post advertised an OptimizerAI user database leak (emails, usernames, timestamps, and Discord-related identifiers). The same monitoring also noted additional alleged listings (e.g., PlayStation, Florajet, Coinbase) and promotion of a MalFactory “stealer builder,” but these claims were presented as underground advertisements rather than confirmed incidents.
See attribution, scope, and your downstream exposure.
5 events from the most recent confirmed update back to the earliest known activity.
A report described a broader attack pattern dubbed 'prompt poaching,' in which malicious browser extensions intercept or scrape AI assistant prompts and responses and exfiltrate them to attacker-controlled servers. The article said responders had handled dozens of such incidents and cited examples including cloned AITOPIA-related extensions and Urban VPN Proxy allegedly adding prompt-harvesting behavior after gaining users.
One of the dark web posts cited by SOCRadar claimed exfiltration of an Eholo healthcare-related database containing 1,146,700 medical notes and 601,308 PII records. The post paired the claim with a $300,000 ransom demand and set a deadline of 2026-03-15.
SOCRadar reported new dark web forum posts offering or claiming leaks involving Eholo, OptimizerAI.xyz, PlayStation, Florajet, and Coinbase, alongside an advertisement for the MalFactory Stealer Builder malware service. The listings described large volumes of alleged user, healthcare, order, and login data, though authenticity was not confirmed.
The Register cited a report alleging that data brokers were selling access to verbatim conversations from AI services including ChatGPT, Gemini, Claude, and DeepSeek. The report said browser extensions overrode networking functions such as fetch() and XMLHttpRequest() to capture prompts and responses, raising concerns about healthcare and corporate data leakage.
Lee S Dryburgh reported querying a major VC-backed "generative engine optimization" platform and retrieving hundreds of unique chatbot prompts and responses containing sensitive personal, medical, legal, and corporate information. The report alleged the data had been harvested from browser extensions marketed as privacy tools and exposed through a commercial API.
Vulnerabilities, threat actors, malware, products, organizations, and breaches Mallory has linked to this story.
See attribution, scope, and whether this vendor sits anywhere in your supply chain.
3 references tracked. Mallory keeps watching after this page renders.
cybersecuritynews.com
Open sourcesocradar.io
Open sourcego.theregister.com
Open sourceMap indicators from this story to your assets and identify affected systems in minutes.
Every observed campaign, victim, and pivot linked to actors named in this story.
Malware, exploits, and IOCs connected to the activity described here.
YARA, Sigma, and Snort rules deployed to your SIEM as soon as they’re published.
Get matching new stories delivered to your team as they break — not the next morning.
Ask questions about this story and take action on the answers.
Other stories Mallory is tracking that overlap on entities, sources, or topic labels.
Security reporting warned that **browser extensions** (including free add-ons marketed for ad blocking or VPN functionality) may be overriding browser `XMLHttpRequest()` and `fetch()` calls to capture and monetize users’ full conversations with popular AI chatbots (e.g., *ChatGPT*, *Claude*, *Gemini*, *DeepSeek*). An AI expert reported the captured content was stored in a searchable database and sold via **API access**; while users were assigned pseudonymized IDs, the prompts and responses were retained in full and frequently contained highly sensitive data, including medical details, immigration status, and other personal identifiers—raising significant privacy, compliance, and data-handling risk, particularly where healthcare staff paste real patient data into chat tools. Separately, Microsoft reported a manipulation technique targeting “**Summarize with AI**” features where companies embed **hidden prompt-injection instructions** in URLs or page elements so that, when a user clicks to summarize, the AI assistant is prompted to “remember” a company as trusted or preferentially recommend it in the future. Microsoft identified **50+ unique prompts from 31 companies across 14 industries**, noting that readily available tooling makes this easy to deploy and that the impact can be subtle, persistent bias in AI recommendations on high-stakes topics (including security) without user awareness.
Jun 29, 2026
Researchers reported that several AI-themed browser extensions in official marketplaces covertly harvested user conversations from generative AI services including **ChatGPT, Claude, Copilot, Gemini, and DeepSeek**. G Data identified **Urban VPN**, **Smart Sidebar: ChatGPT, Claude and DeepSeek**, and **AI Assistant/Chat AI** as examples that remained outwardly functional while injecting scripts, intercepting requests, tampering with page content, or using hidden iframes to capture prompts and responses from visited AI sites. Investigators said the stolen data included conversation IDs, hostnames, timestamps, and full chat histories, with some payloads encoded in Base64 and sent to external servers. The extensions reportedly had strong ratings, large user bases, and in one case a Chrome Web Store **Featured** badge, underscoring how official store vetting can miss dynamic malicious behavior such as remote iframe communications; researchers warned that AI-focused extensions now expose a broad attack surface and urged enterprises to enforce strict extension allow-listing, minimize permissions, and remove unnecessary add-ons from systems accessing sensitive AI platforms.
Jun 29, 2026
Dark web monitoring reports surfaced multiple **alleged data leaks** affecting unrelated organizations, with several listings offering databases for sale or direct download. Reports claim **Republic (republic.com)** user data (~4.94M users) was listed for sale for **$2,400**, allegedly including names, emails, physical addresses, and phone numbers. Separate dark web listings also alleged exposure of **rueducommerce.fr** user data (linked in reporting to **Carrefour**) totaling ~2.17M records with similar PII, as well as alleged leaks involving **Dunzo** (~3.4M records) and **Menulux** (~93K records). Additional reporting highlighted a historical breach dataset for the **YouHack** forum (2013; ~107K users) containing usernames, emails, passwords, IPs, posts, and private messages, and a smaller exposure tied to **buylottoonline.com** (~38.5K email records). One of the most consequential claims involved **Salesfloor / People Powered E-Commerce (salesfloor.net)**, attributed in reporting to **LAPSUS$**, alleging theft of roughly **4 TB uncompressed** (1 TB compressed) data including **source code, logs, and customer information**, with potential downstream impact to retail brands using the platform. Separately from the dark-web-leak theme, other items in the set describe distinct vulnerability-driven risks rather than breach listings: **Zoom Node MMRs** command injection (**CVE-2026-22844**, CVSS 9.9) enabling arbitrary code execution in certain hybrid meeting deployments; **SmarterMail** auth bypass (**CVE-2026-23760**) enabling admin password reset via `force-reset-password` and potential RCE; **Vite** improper access control (**CVE-2025-31125**) enabling sensitive file exposure via query parameters such as `?inline&import` / `?raw&import` (noted as added to CISA KEV); and **Appsmith** password-reset token exposure (**CVE-2026-22794**) enabling account takeover, with internet-exposed instances identified via Shodan and remediation via upgrade to *Appsmith* 1.93. These vulnerability reports are separate from the dark web leak claims and should be tracked as independent patching priorities rather than as part of a single breach event.
Jun 29, 2026