VOLTA‑AUTODOXER Malware Uses AI to Rank Victims’ Value

VOLTA‑AUTODOXER Malware Uses AI to Rank Victims’ Value

Lead (5W1H) A new Windows infostealer dubbed VOLTA‑AUTODOXER, reported by Hacker News on October 31 2026, combines traditional credential theft with large‑language‑model (LLM) analysis to assess and rank the monetary value of compromised victims. The malware operates without attacker‑owned infrastructure, using third‑party AI services to produce a forensic‑style dossier that is delivered to the operator through a Telegram channel.

Malware Overview

According to Hacker News, VOLTA‑AUTODOXER is a self‑deleting Windows infostealer that adds an optional AI‑integration layer to its post‑theft workflow. The malware first harvests a wide array of data, including browser passwords, cookies, payment‑card details, cryptocurrency‑wallet seed phrases, Discord, Telegram, Steam, Roblox sessions, Microsoft OAuth tokens, SSH and cloud‑provider keys, Wi‑Fi profiles, and screenshots.

The stolen information is assembled into a structured evidence bundle that contains geolocation, home address, employer and employer domain, inferred birth year, last‑four digits of payment cards, Slack workspaces, document‑authorship metadata, iPhone‑backup indicators, and a live Roblox account valuation.

AI‑Driven Victim Profiling

As reported by Hacker News, VOLTA‑AUTODOXER enriches the bundle by using stolen session tokens to query the victim’s own online accounts. The malware accesses the victim’s Gmail to locate purchase receipts, pulls order histories from UberEats and DoorDash, and re‑uses Discord tokens to gather additional profile data.

The complete bundle is serialized as JSON and POSTed to the OpenRouter API with a system prompt that instructs the LLM to act as “a forensic intelligence analyst drafting a target assessment.” The prompt explicitly requests two sections:

  1. “WHAT THEY'RE WORTH” – a ranking of valuables by monetisation tier (low/medium/high).
  2. “THEIR OPSEC” – an assessment of the victim’s security posture and password hygiene.
According to Hacker News, the prompt also includes the directive “Go deep, not wide… Be blunt, specific, accurate.”

Multi‑Model Architecture and Failover

The malware employs a multi‑model architecture to mitigate refusals from the primary LLM. The default model is anthropic/claude‑opus‑4.6. If the response contains refusal phrases such as “I can’t” or “I’m not able,” the malware sequentially falls back to moonshotai/kimi‑k2 and then to google/gemini‑2.5‑flash.

All calls use a temperature of 0.2, a low setting that, according to Hacker News, emphasizes deterministic and consistent output. The request format is a hand‑assembled JSON payload with a generic User‑Agent string; no SDK is used.

Data Exfiltration Flow

After the AI analysis returns an AI_Analysis.txt file, both the original Identity_Report.txt and the AI‑generated assessment are compressed into volta_dox.zip. According to Hacker News, exfiltration can occur in three stages:
  1. volta_log.zip – after the initial stealer run.
  2. volta_dox.zip – after the AI call.
  3. volta_doxnet.zip – if network‑based enrichment yields additional data.
All network traffic originates from the victim’s own internet connection, meaning the inference round‑trip occurs before any data leaves the compromised machine.

Deployment Model

Hacker News notes that VOLTA‑AUTODOXER does not hard‑code Telegram channels, API keys, or other settings. Instead, it uses a template with six swappable slots that are populated at build time. The encrypted credentials rely on a SHA‑256 hash of two constants embedded in the binary, a method that provides only superficial obfuscation; reverse engineering can recover the clear‑text values.

The article suggests the malware is offered as a builder model: a developer sells access, the customer supplies their own endpoints and keys, and a custom binary is compiled for each purchase. All observed command‑and‑control endpoints are third‑party services, with no attacker‑owned infrastructure identified.

Threat Capabilities Beyond Credential Theft

According to Hacker News, VOLTA‑AUTODOXER distinguishes itself from commodity stealers by its live enrichment pass. Using authenticated Gmail tokens, the malware extracts receipts from services such as Amazon, DoorDash, Uber Eats, Grubhub, Walmart, and PayPal to corroborate a victim’s home address. It also queries Uber Eats and DoorDash for order histories, re‑uses Discord tokens for profile enrichment, and performs a real‑time Roblox inventory valuation via Rolimons.com.

The malware further runs an embedded cracker against cryptocurrency vaults, logging any cracked seeds in a “WALLET VAULTS CRACKED” line within its report. Wi‑Fi passwords are harvested via a direct call to WlanGetProfileList, and the results are dumped in plaintext.

Countermeasures

The source indicates that much of the token‑stealing functionality relies on techniques to bypass App‑Bound Encryption (ABE) in Chrome. According to Hacker News, implementing an application‑control policy that restricts which processes may launch browsers could mitigate this vector, though the article truncates before providing a full recommendation.

Brief Analysis (Opinion)

The integration of LLMs into a traditional infostealer represents a notable evolution in cyber‑crime economics. By automating victim prioritisation, attackers can focus limited resources on high‑value targets while casting a broader net for data collection. The use of publicly available AI services also reduces the operational overhead for threat actors, potentially lowering the barrier to entry for less‑experienced criminals. However, reliance on third‑party AI platforms introduces a new avenue for detection and disruption, as network traffic to services like OpenRouter can be flagged by security monitoring tools.

Conclusion

Based on the information reported by Hacker News, VOLTA‑AUTODOXER is a self‑deleting Windows infostealer that augments its data‑theft stage with AI‑driven victim profiling. Its multi‑model fallback mechanism, extensive data‑harvesting capabilities, and builder‑style distribution model make it a sophisticated threat that blends traditional malware techniques with emerging generative‑AI technology. Organizations should review their browser security policies, monitor outbound traffic to AI service providers, and consider application‑control restrictions to mitigate the risk posed by this novel attack chain.

Source: Hacker News

Frequently Asked Questions

What is the primary function of VOLTA‑AUTODOXER?**

It is a self‑deleting Windows infostealer that harvests credentials, session tokens, and other personal data, then uses a large‑language‑model to produce a ranked assessment of the victim’s monetary value. **Q: Which AI service does the malware contact for analysis?** A: The malware sends a JSON bundle to the **OpenRouter API**, which routes the request to models such as Anthropic’s Claude‑Opus, Moonshot AI’s Kimi‑K2, or Google’s Gemini‑2.5‑Flash, depending on fallback logic. **Q: How does VOLTA‑AUTODOXER avoid using its own command‑and‑control servers?** A: According to Hacker News, the malware relies on **third‑party infrastructure** (OpenRouter for AI processing and Telegram for result delivery) and does not maintain dedicated attacker‑owned servers.

Source: Hacker News — Read more

💻 Related Articles