Introduction and Project Genesis
The speaker, Alexer from GetWatcher (an NDR security solution focused on threat hunting and analysis), introduces the talk on French stealer groups. The project originated in 2024 amid a wave of undocumented cyber attacks targeting French streamers, media, and major companies. These groups were small, discrete, and lacked documentation, making them hard to track but highly dangerous. The goal was to document them, starting with Nova Stealer (not to be confused with other Nova groups) as a representative example. Thanks are given to former intern Nicholas for the initial idea. The presentation covers the stealer malware, the groups, lack of awareness, and the need for close monitoring.
Nova Stealer Overview and Evolution
Nova Stealer is a Go-based malware (e.g., "GS.pike") that evolved from basic, non-obfuscated versions to more sophisticated, obfuscated ones (though still reversible). Key evolutions include:
- Panel and Interface: Started simple (left screenshot in presentation) and became more complex with better UI for data storage and features.
- Targeting Shifts: Early versions targeted broad items like "Pron" (a student-professor communication tool), but later focused on cryptocurrency, bank accounts, and personal data.
- Rebranding and Iterations: The group (likely the same person or small team) frequently rebranded with minor updates, adding obfuscation or new data handling methods. This evolution mirrors broader cybercrime trends, where small groups mature if given time, moving from unsophisticated to more capable threats.
How the Stealer Works
The malware operates as a builder-as-a-service:
- Acquisition: Users buy a key/token via a selling website, then configure via Telegram or Discord bots.
- Building and Delivery: Generates an executable (.exe) file, often disguised as fake video streams, game cheats, or online tools targeting individuals (not corporates directly).
- Exfiltration: Data is sent to the buyer's Discord initially; later versions added collection websites and third-party services like GoFile. Features a "dual hook" where both the buyer and the group receive stolen data.
- Customization: Later versions (e.g., Light Nova) allow tailored environments for data use, often with fake or low-value info for demos. Similarities were found with open-source stealers on GitHub, suggesting heavy copy-pasting of code, spreading techniques, and infrastructure (e.g., Discord integration).
Group Structure and Mapping
The ecosystem involves multiple interconnected groups and individuals:
- Mapping: Blue nodes represent individuals (possibly physical persons), pink for groups. Many actors split time across groups.
- Hierarchy: Often self-disclosed.
- Developers/Owners: Handle malware evolution, updates, and advertising.
- Community Managers: Manage Telegram/Discord communities; recreate servers if shut down.
- Support: Provides user assistance, mainly on Telegram. Groups are small (15-25 years old, often students/university attendees), with recruitment ads noting exam periods or diplomas. One group freaked out on LinkedIn after a report, denying involvement (proven false). Admins take "vacancies" like holidays, sharing identifiable photos (e.g., airport seats).
OSINT Techniques Breakdown
The speaker details OSINT (Open Source Intelligence) methods used to profile and track these groups, leveraging public/self-shared data. Broken down into key sections:
Self-Doxing and Internal Conflicts
- Groups/admins often dox each other during disputes (e.g., at 3 AM in Telegram chats).
- Method: Monitor Telegram messages before deletion; log suppressed content for names, addresses, or meetup spots (e.g., "Go to this address to meet this guy").
- Insights: Revealed hierarchies, personal details, and connections; one resume was found (not shared due to sensitivity).
Shared Language and Recruitment Analysis
- Analyzed shared slang, language patterns, and recruitment posts (e.g., Nova's form asking about exam periods/diplomas).
- Method: Cross-reference across groups via ads linking channels; progress from one group to another by following invitations.
- Insights: Confirmed age range (15-25, students); some groups require buying the stealer to stay, leading to self-reveals (e.g., "I'm part of these other groups").
Personal Sharing and Visual Analysis
- Admins share personal photos (e.g., airport seats, plane models) during "vacancies."
- Method: Geotag/timestamp analysis; cross-reference with flight data (e.g., model, seat, time to infer flight; hypothetically access boarding lists via law enforcement for full ID).
- Insights: Identified travel details; combined with resumes for full profiles.
Forum and Dark Web Digging
- Tracked sales on forums like XSS, RaidForums (e.g., stolen data from companies via Epsilon group).
- Method: Search for identical messages across forums; analyze "proofs" provided (e.g., screenshots with Telegram usernames); dig into usernames for aliases/connections.
- Insights: Linked to scams (e.g., buying Revolut accounts for potential money laundering); discovered drug sales (cocaine via Telegram bot "Walls," tied to arrests in Gabon per French press).
Code and Infrastructure Analysis
- Retrieved Nova's source code; compared with other GitHub stealers.
- Method: Check for common techniques (e.g., Discord hooks, exfiltration paths); map shared infrastructure.
- Insights: Confirmed copy-paste ecosystem; identified similar malwares.
Community Access and Monitoring
- Channels are often open/public.
- Method: Start in one group, follow ads to others; no payment needed initially (though sometimes prompted).
- Insights: Built the full group map; monitored for passwords (rare; mostly game accounts like Roblox, Battle.net for skins/resales).
Additional Activities and Diversification
Beyond stealers, groups engage in:
- Illegal Content: Selling leaks, managing revenge porn, sextortion.
- DDoS Services: C2 networks for zombie attacks, advertised in stealer communities.
- Data Reselling: Via forums (e.g., Epsilon's company data sales); dual hooks enable group access.
- Broader Crime: Links to drug networks (e.g., "Walls" Telegram for cocaine); Epsilon case study showed forum overlaps proving connections. This shows young cybercriminals diversifying, using cyber tools for offline crimes.
Importance of Monitoring
Small groups pose big risks:
- Indirect Corporate Threats: Personal devices may hold work data; kids downloading fakes can expose company info.
- Vulnerabilities: Password reuse between personal/professional; data enters combo lists for larger attacks.
- Broader Impact: Users (e.g., €4/week buyers) underestimate consequences; actions enable further crimes without realizing illegality/danger. Monitoring covers overlooked spaces; reports on Nova and others provide deeper insights.
Technical Insights and Case Studies
- Profiling: Combined OSINT for actor links.
- Epsilon Case: Linked stealer sales to forum scams/drugs via username cross-referencing. Groups need ongoing monitoring despite seeming "dumb" (e.g., leaving tracks).
Q&A Summary
- Password Monitoring: Rarely steal passwords; focus on game accounts (Roblox, Ubisoft) for skins/resales, or Discord Nitro.
- Channel Access: Open via ads; chain from one group to another (no initial payment; sometimes prompted to buy/test).
- Info Sharing with Law Enforcement: Yes, shared due to value; no point in hoarding. No further questions noted.