Varonis reports TheMailer phishing kit uses attacker-chosen AI models
Dark Reading reports Varonis research says TheMailer phishing kit uses LLMs to tailor lures; review email authentication and link checking.
- Published
- Ecosystem
- AI developer tools
- Exploitation
- Not confirmed. Dark Reading says it is unclear how successful the kit is, and Varonis assesses it has a large customer base.
What happened
According to Dark Reading, new Varonis research describes “TheMailer,” a commercial phishing kit that uses large language models (LLMs) to automate much of the phishing process, including tailoring email lures to get past content filters (Dark Reading). These are claims from Varonis as relayed by one outlet. Dark Reading says it is unclear how successful the kit is, and no independent verification is given in the source.
Per Dark Reading’s account of the Varonis blog post:
- A customer buys a lifetime license for between $3,000 and $6,000 in cryptocurrency, depending on the deal and which AI functions are included.
- The operator supplies their own API key and picks a model. Claude, ChatGPT, DeepSeek, Grok, Gemini and Mistral are listed as included.
- Customers reportedly only need to bring SMTP servers, proxies and target email lists.
- The kit can reportedly generate lures from a short prompt or a real brand email, personalize each email (address, body, subject line), remove AI writing patterns, and adjust content for spam filters. Varonis researcher Daniel Kelley says it can tailor lures to each mail provider and “rewrite them until they bypass spam filters.”
Kelley told Dark Reading that the seller’s screenshots show a single campaign of nearly 25,000 emails and a list of more than 145,000 targets sorted by mail provider. He said every recipient gets their own version of the email, subject line and attachment. He does not call it a spear-phishing tool: the personalization seen is at the level of a name on an invoice, and Varonis did not see features for researching individual targets. He described it as opportunistic phishing with the polish of something targeted (Dark Reading).
Not confirmed in the source: the number of customers (Varonis only assesses it as large), any named victims, any named actor behind the kit, and the real-world success rate of campaigns. No CVE or software vulnerability is involved.
Am I affected?
This is a phishing tool, not a flaw in a product, so there is nothing to patch. Any organization that receives email could in principle be a target. The source does not name any victim or sector. Kelley says he expects more kits to follow, “because once one seller offers this, buyers start asking for it elsewhere” (Dark Reading).
What to do now
The source attributes these recommendations to Varonis:
- Consider enforcing email authentication: SPF, DKIM and DMARC.
- Monitor for lookalike and new domains.
- Check links through full redirect chains at the time of the click.
- Update phishing training for a post-AI world. Varonis says bad grammar is no longer a reliable sign, so teach users to be wary of urgency, unexpected requests and login prompts, especially “unusual activity” or account statement alerts.
- Apply least-privilege access and data monitoring to limit damage if someone clicks anyway.