Skip to main content

Agentic URL Scanning 2.0 Integration

  • July 30, 2026
  • 2 replies
  • 34 views

mikemeeks
Staff
Forum|alt.badge.img+5

 

Agentic now leverages VirusTotal's sandboxed browser execution to analyze suspicious URLs dynamically.

Capabilities:

  • Runtime Telemetry: Captures dynamic DOM modifications, network requests, and global JavaScript variables.
  • Campaign Mapping: Fingerprints phishing kits (e.g. pivoting on __jwrAuth variables) to map related threat infrastructure.

URLs that are not in the public corpus of Google Threat Intelligence are analyzed privately using Private Scanning.

Example Prompt: "Investigate hxxps://t-mobile.zyhqxp.top/pay/ and find other URLs from the same campaign using its JS footprint. Then create a campaign collection."
View this conversation in Agentic →

 

What This Means for You

As security analysts, we've all been trapped in the endless game of "whack-a-mole": you block one malicious URL, and the threat actor spins up three more before you can even close the ticket. Historically, dissecting these evasive campaigns required spinning up local sandbox environments, manually parsing raw code, and hoping you didn't miss a dynamic redirect.

With Agentic URL Scanning 2.0, We are changing the rules of the game.
By combining Google Threat Intelligence (GTI) with VirusTotal's sandbox browser detonation, our AI agents don’t just read code—they dynamically interact with it. Here is what this looks like in your daily operations:

Stop chasing single links; block the whole campaign: Instead of chasing individual indicators, you can pivot on dynamic fingerprints (like unique JavaScript variables) to unmask the attacker's entire infrastructure in seconds.

Analyze in complete safety and privacy: You get deep, live telemetry on suspicious URLs—including Tor-hosted onion domains—privately. Your investigations remain hidden from threat actors without you having to build or manage an isolated malware lab.

Accelerate time-to-insight: What used to take hours of manual reverse-engineering is now accomplished with a simple, conversational prompt.

The Bottom Line: This upgrade shifts your security operations from reactive triaging to proactive threat demolition, giving your defenders the speed and context needed to outsmart modern threat actors.

2 replies

whathehack81
Forum|alt.badge.img+7

This is a solid capability, especially for URLs that are not yet present in the public corpus.
How is private scanning data handled? Specifically:
Are submitted URLs, DOM content, and network requests retained?
Can private scan results be restricted to the customer tenant?
Does campaign mapping expose the underlying runtime indicators and evidence used to group URLs?
Being able to review that evidence would make the output much easier to validate before using it in detection or response workflows.


mikemeeks
Staff
Forum|alt.badge.img+5
  • Author
  • Staff
  • July 30, 2026

This is a solid capability, especially for URLs that are not yet present in the public corpus.
How is private scanning data handled? Specifically:
Are submitted URLs, DOM content, and network requests retained?
Can private scan results be restricted to the customer tenant?
Does campaign mapping expose the underlying runtime indicators and evidence used to group URLs?
Being able to review that evidence would make the output much easier to validate before using it in detection or response workflows.


Thanks for these fantastic questions! You are spot on—having full visibility and being able to independently validate the underlying evidence is absolutely critical before importing threat intelligence into active detection or automated response workflows. When you utilize Private Scanning, all submitted URLs, dynamic DOM content, and network telemetry are kept strictly confidential—they are detonated in an isolated sandbox, stored temporarily in secure private buckets, and permanently deleted upon completion of the analysis. These scan results are logically restricted entirely to your authorized customer tenant group, ensuring your active investigations remain completely private and are never exposed to the public VirusTotal corpus, other organizations, or threat actors. Furthermore, campaign mapping is completely transparent and does not operate as a "black box," fully exposing the underlying runtime indicators (such as specific JavaScript variables, redirect chains, or PCAP data) so that your security team can easily inspect, validate, and extract the evidence before deploying high-fidelity rules in your active detection and response workflows.