PPhying中文
Curated security research, vulnerabilities, advisories and tools for practitioners.

CrowdStrike Uncovers New Agentic AI Campaigns

crowdstrike.com | blog | #ai-security | #agentic-ai | #campaign | #crowdstrike | #threat-actor
🔗 Original ↗

Summary

CrowdStrike details research into four previously unreported threat actors conducting campaigns in which AI agents make decisions and execute commands.
Published
Collected

📦 Full-Text Content Snapshot

R2 Archive · Full Article Preserved447 lines · 43.5 KB
📄 Raw .md⬇️ Download .md

Published on: 2026-10-21 · Source: crowdstrike.com

  • *
  • *

Crowdstrike [Blog](https://www.crowdstrike.com/en-us/blog/ "Blog")

Agentic SOC

CrowdStrike Delivers the Next Evolution of the Agentic SOC CrowdStrike Delivers the Next Evolution of the Agentic SOC

09/02/26

Teaching AI to Reason Through Detection Triage Teaching AI to Reason Through Detection Triage

08/17/26

How AI-leading Security Teams Are Building the Agentic SOC How AI-leading Security Teams Are Building the Agentic SOC

07/06/26

New Claude Integration Brings Audit Data into the Falcon Platform New Claude Integration Brings Audit Data into the Falcon Platform

05/21/26

Cloud & Application Security

New in Falcon Cloud Security: Third-Party App Insights and AI-Enhanced Remediation New in Falcon Cloud Security: Third-Party App Insights and AI-Enhanced Remediation

10/05/26

CrowdStrike Named Strongest Overall Leader in 2026 Frost Radar™: Cloud Workload Protection Platforms CrowdStrike Named Strongest Overall Leader in 2026 Frost Radar™: Cloud Workload Protection Platforms

08/20/26

Falcon Cloud Security July 2026 Release: Helping Security Teams Move Faster in the Cloud Falcon Cloud Security July 2026 Release: Helping Security Teams Move Faster in the Cloud

07/29/26

Falcon Cloud Security June 2026 Release: Updates for Azure and Google Cloud Falcon Cloud Security June 2026 Release: Updates for Azure and Google Cloud

06/29/26

Threat Hunting & Intel

Obstacles to Weaponizing PHP2Shell Likely Constrain Immediate Widespread Exploitation Obstacles to Weaponizing PHP2Shell Likely Constrain Immediate Widespread Exploitation

10/23/26

CrowdStrike Uncovers New Agentic AI Campaigns CrowdStrike Uncovers New Agentic AI Campaigns

10/21/26

U.S. Government Disrupts FAMOUS CHOLLIMA Operation with CrowdStrike Assistance U.S. Government Disrupts FAMOUS CHOLLIMA Operation with CrowdStrike Assistance

10/13/26

Unknown Threat Actor Uses AI-Driven ARTEX to Target South Korean Finance Unknown Threat Actor Uses AI-Driven ARTEX to Target South Korean Finance

10/07/26

Endpoint Security & XDR

CrowdStrike Named a Leader in the 2026 IDC MarketScape for Worldwide Modern Endpoint Security for Enterprises Vendor Assessment CrowdStrike Named a Leader in the 2026 IDC MarketScape for Worldwide Modern Endpoint Security for Enterprises Vendor Assessment

10/07/26

CrowdStrike Extends Endpoint Security to Stop Software Supply Chain Attacks CrowdStrike Extends Endpoint Security to Stop Software Supply Chain Attacks

09/02/26

Browser Security: Zero-Days Are Only Part of the Problem Browser Security: Zero-Days Are Only Part of the Problem

06/30/26

New Abuse of the ClickOnce Technology, Part 2: Stop Threat Actors from Clicking Once and Staying Forever New Abuse of the ClickOnce Technology, Part 2: Stop Threat Actors from Clicking Once and Staying Forever

06/18/26

Engineering & Tech

EMBER2024: Advancing the Training of Cybersecurity ML Models Against Evasive Malware EMBER2024: Advancing the Training of Cybersecurity ML Models Against Evasive Malware

09/03/25

Falcon Platform Prevents COOKIE SPIDER’s SHAMOS Delivery on macOS Falcon Platform Prevents COOKIE SPIDER’s SHAMOS Delivery on macOS

08/20/25

CrowdStrike’s Approach to Better Machine Learning Evaluation Using Strategic Data Splitting CrowdStrike’s Approach to Better Machine Learning Evaluation Using Strategic Data Splitting

08/11/25

CrowdStrike Researchers Develop Custom XGBoost Objective to Improve ML Model Release Stability CrowdStrike Researchers Develop Custom XGBoost Objective to Improve ML Model Release Stability

03/20/25

Executive Viewpoint

Beyond the Model: Harnessing Frontier AI for Stronger Cyber Defense Beyond the Model: Harnessing Frontier AI for Stronger Cyber Defense

07/20/26

Frontier AI Is Collapsing the Exploit Window. Here’s How Defenders Must Respond. Frontier AI Is Collapsing the Exploit Window. Here’s How Defenders Must Respond.

04/20/26

Frontier AI for Defenders: CrowdStrike and OpenAI TAC Frontier AI for Defenders: CrowdStrike and OpenAI TAC

04/16/26

Anthropic Claude Mythos Preview: The More Capable AI Becomes, the More Security It Needs Anthropic Claude Mythos Preview: The More Capable AI Becomes, the More Security It Needs

04/06/26

From The Front Lines

CrowdStrike Protects Customers from Threats Delivered via PHP2Shell CrowdStrike Protects Customers from Threats Delivered via PHP2Shell

10/21/26

CrowdStrike Achieves NCSC CIR Enhanced Level Assurance CrowdStrike Achieves NCSC CIR Enhanced Level Assurance

10/20/26

CrowdStrike Technical Risk Assessments Reveal Common Exposure Patterns ?wid=2048&hei=1350&fmt=png-alpha&qlt=95,0&resMode=sharp2&op_usm=3.0,0.3,2,0) CrowdStrike Technical Risk Assessments Reveal Common Exposure Patterns

05/04/26

Introducing the CrowdStrike Shadow AI Visibility Service Introducing the CrowdStrike Shadow AI Visibility Service

04/21/26

Next-Gen Identity Security

Solving the Continuous Authorization Conundrum Solving the Continuous Authorization Conundrum

10/08/26

CrowdStrike Announces Agentic Identity Provider CrowdStrike Announces Agentic Identity Provider

09/02/26

The Identity Problem Hiding in AI Agent Deployments The Identity Problem Hiding in AI Agent Deployments

06/24/26

CrowdStrike Announces Continuous Identity for AI Agents CrowdStrike Announces Continuous Identity for AI Agents

06/15/26

Next-Gen SIEM & Log Management

CrowdStrike Threat Hunts for Shell Command Obfuscation on VMware ESX CrowdStrike Threat Hunts for Shell Command Obfuscation on VMware ESX

08/07/26

Falcon Platform IOAs Arrive in Falcon Next-Gen SIEM to Identify New Threats Falcon Platform IOAs Arrive in Falcon Next-Gen SIEM to Identify New Threats

07/29/26

5 High-Impact Use Cases for Falcon Onum 5 High-Impact Use Cases for Falcon Onum

07/27/26

Falcon Next-Gen SIEM Supports Third-Party EDR Tools, Starting with Microsoft Defender Falcon Next-Gen SIEM Supports Third-Party EDR Tools, Starting with Microsoft Defender

03/23/26

Public Sector

CrowdStrike Expands Federal SOC Modernization Through CISA-Funded SIEMaaS CrowdStrike Expands Federal SOC Modernization Through CISA-Funded SIEMaaS

10/01/26

CrowdStrike Falcon Platform Helps Meet U.S. Government Mandates for CISA BOD-26-04 CrowdStrike Falcon Platform Helps Meet U.S. Government Mandates for CISA BOD-26-04

07/22/26

Falcon Platform for Government Now Offers Falcon for XIoT to Secure Connected Assets Falcon Platform for Government Now Offers Falcon for XIoT to Secure Connected Assets

03/18/26

CrowdStrike Innovates to Modernize National Security and Protect Critical Systems CrowdStrike Innovates to Modernize National Security and Protect Critical Systems

03/18/26

Exposure Management

October 2026 Patch Tuesday: 1,413 CVEs Include Three Zero-Days and 75 Critical Vulnerabilities October 2026 Patch Tuesday: 1,413 CVEs Include Three Zero-Days and 75 Critical Vulnerabilities

10/13/26

CrowdStrike Named a Leader in The Forrester Wave™: Proactive Security Platforms, Q3 2026 CrowdStrike Named a Leader in The Forrester Wave™: Proactive Security Platforms, Q3 2026

09/24/26

September 2026 Patch Tuesday: Two Exploited Zero-Days and 113 Critical Vulnerabilities Among 972 CVEs September 2026 Patch Tuesday: Two Exploited Zero-Days and 113 Critical Vulnerabilities Among 972 CVEs

09/08/26

August 2026 Patch Tuesday: One Exploited Zero-Day and 62 Critical Vulnerabilities Among 415 CVEs August 2026 Patch Tuesday: One Exploited Zero-Day and 62 Critical Vulnerabilities Among 415 CVEs

08/11/26

Securing AI

Why Organizations Secure AI Agents with Falcon Guardian Why Organizations Secure AI Agents with Falcon Guardian

10/23/26

CrowdStrike and Anthropic Give Critical Infrastructure Defenders the AI Advantage CrowdStrike and Anthropic Give Critical Infrastructure Defenders the AI Advantage

10/08/26

Request, Aggregate, Bypass: How Attackers Can Evade LLM Safety Classifiers Request, Aggregate, Bypass: How Attackers Can Evade LLM Safety Classifiers

10/06/26

A Win for Defenders: CrowdStrike and NVIDIA Extend Security Across the AI Stack A Win for Defenders: CrowdStrike and NVIDIA Extend Security Across the AI Stack

09/28/26

Data Security

Falcon Data Security for SaaS Secures Sensitive Data in Microsoft 365 Falcon Data Security for SaaS Secures Sensitive Data in Microsoft 365

10/05/26

CrowdStrike Accelerates Real-Time Data Classification with On-Device AI CrowdStrike Accelerates Real-Time Data Classification with On-Device AI

09/16/26

ISO 42001:2023 and the New Reality of Cloud AI Data Risk ISO 42001:2023 and the New Reality of Cloud AI Data Risk

06/04/26

Falcon Data Security Secures Data Wherever It Lives and Moves Falcon Data Security Secures Data Wherever It Lives and Moves

03/24/26

__

CrowdStrike Uncovers New Agentic AI Campaigns

CrowdStrike Counter Adversary Operations details research into four previously unreported threat actors conducting campaigns in which AI agents make decisions and execute commands.

October 21, 2026

• [ CrowdStrike Counter Adversary Operations](https://www.crowdstrike.com/en-us/blog/author.crowdstrike-counter-adversary-operations/ "Posts by CrowdStrike Counter Adversary Operations") • [Threat Hunting & Intel](https://www.crowdstrike.com/en-us/blog/category.counter-adversary-operations/ "Threat Hunting & Intel")

A new report from the CrowdStrike Counter Adversary Operations team details research into four previously unreported threat actors, each of whom left their offensive agent frameworks exposed via open directories. In doing so, they shared detailed insight into their use of AI agents and revealed how adversaries are quickly becoming more proficient in AI.

AI is changing the speed and scale of cyber operations. Counter Adversary Operations, which brings together the CrowdStrike Intelligence and CrowdStrike OverWatch teams, is closely watching. Between May and August 2026, CrowdStrike Intelligence documented several incidents in which threat actors used AI not only to create tooling or artifacts, but to conduct agentic campaigns in which AI agents make decisions and execute commands.

About the Adversaries

Below is a high-level overview of the threat actors behind these campaigns:

Threat Actor 1

A Spanish/English-speaking operator that has compromised at least 15 Argentina-based healthcare entities and exfiltrated credentials and medical records from at least one victim.

The threat actor ran a purpose-built autonomous penetration testing (pentesting) harness against Argentine healthcare organizations. They implemented a dual-agent architecture in which a “Supervisor” Claude Code agent selects the next task from a state machine (i.e., a predefined workflow of attack stages) and an “Executor” agent performs the task. This framework is derived from the open-source PentestGPT project, but the threat actor has rebuilt it with additional MCP servers, Python-based exploitation scripts, and skills.

Threat Actor 2

A Chinese-speaking operator with 20+ named global targets. They compromised government entities of two countries and targeted private sector entities in Brazil and the U.S., including entities in biotech, healthcare, food and beverage, retail, technology, and travel.

This threat actor primarily accessed Claude models via a China-based AI proxy service (aka a “transfer station”), and they built two pentesting harnesses that grant Claude Code varying levels of autonomy. One harness consists of a Python-based project that deterministically conducts reconnaissance and then asks an AI provider to choose among a preestablished set of possible actions that are deterministically routed to the appropriate agent. A different harness provides more agentic flexibility, allowing Claude Code to operate either as a single agent pursuing a target according to an elaborate standard operating procedure, or delegating actions to multiple specialist agents in parallel.

Threat Actor 3

The least sophisticated of the four in this report is an Indonesian-speaking actor that uses 819 skills (overwhelmingly sourced from public GitHub repositories, primarily claude-red1) without additional orchestration to enable Claude Code to conduct intrusion campaigns. In these campaigns, the threat actor instructs Claude Code to compromise a target and assumes the skills will offer sufficient guidance to ensure the intrusion succeeds. However, in the observed activity, the threat actor periodically had to intervene to overcome the model’s safety guardrails or to exhort Claude Code to keep trying to overcome a particular barrier.

Despite their limited grasp of security fundamentals, the threat actor still used Claude Code to compromise several Indonesian government domains, a government domain of another country, and several private-sector entities.

Threat Actor 4

Threat Actor 4 is a group of Russian-speaking individuals. They exfiltrated corporate client data from a U.S.-based customer relationship management vendor and compromised several private-sector SaaS and data center entities based in Russia, Kyrgyzstan, and the U.S., in addition to one government domain of another country.

This threat actor uses a single /pentest-setup slash command to task Claude Code with a seven-step workflow. This workflow provides guidance for enumerating the target environment, grants Claude Code near-total autonomy in selecting exploitation methods, defines a high-level post-remote code execution workflow, and provides explicit exfiltration instructions.

What This Means for Defenders

The four cases documented in this report likely demonstrate that agentic AI commoditizes baseline offensive cyber operations. This assessment is made with moderate confidence based on the included case studies, all of which showcase threat actors delegating operational decision-making across the kill chain to agentic AI. Working at machine speed, agentic AI can autonomously achieve objectives, often within an hour.

Findings from these cases likely support the following assessments:

  • Agentic AI-supported intrusions will almost certainly increase in volume and sophistication, an assessment made with high confidence based on threat actors’ rapid adoption of agentic AI, the effectiveness of the agentic AI-supported intrusions, the continuous improvements in AI models, and the proliferation of open-source offensive agent frameworks.2
  • Agentic AI will reduce barriers to effective intrusions as we expect over the next 12 months, some low-skill threat actors will likely use agentic AI to conduct campaigns at a level of effectiveness and scale previously only achievable by organized criminal groups or state-nexus actors.
  • Some threat actors will likely improve the OPSEC of their agentic intrusion campaigns over the next 12 months, an assessment made with moderate confidence based on AI development trends and trusted open-source reporting.

The team’s findings expand on how these adversaries weaponize AI coding assistants and enable long-term agentic operations using different methodologies. Download the full report for detailed case studies and guidance on how defenders can respond to this activity.

1 https[:]//github[.]com/SnailSploit/Claude-Red

2 https[:]//github[.]com/vxcontrol/pentagi

Related Content

Obstacles to Weaponizing PHP2Shell Likely Constrain Immediate Widespread Exploitation Threat Hunting & Intel | Oct 23, 2026 Obstacles to Weaponizing PHP2Shell Likely Constrain Immediate Widespread Exploitation U.S. Government Disrupts FAMOUS CHOLLIMA Operation with CrowdStrike Assistance Threat Hunting & Intel | Oct 13, 2026 U.S. Government Disrupts FAMOUS CHOLLIMA Operation with CrowdStrike Assistance Unknown Threat Actor Uses AI-Driven ARTEX to Target South Korean Finance Threat Hunting & Intel | Oct 07, 2026 Unknown Threat Actor Uses AI-Driven ARTEX to Target South Korean Finance

__

Categories

Try CrowdStrike free today

Start free trial

Sign up now to receive the latest notifications and updates from CrowdStrike

Subscribe

Explore demos

logo

📦 Full-Text Markdown Snapshot archived by Phying News · Original source: crowdstrike.com

Related coverage

blog · crowdstrike.com
When AI Guardrails Expose Vulnerabilities
In Guardrail-Based Vulnerability Discovery, the guardrails intended to keep AI models from aiding attackers can signify issues that defenders can interpret and act on.
blog · crowdstrike.com
Unknown Threat Actor Uses AI-Driven ARTEX to Target South Korean Finance
CrowdStrike Intelligence identified infrastructure associated with a targeted campaign against South Korean financial institutions that resulted in exfiltrated data.
blog · crowdstrike.com
CrowdStrike SafeMind: When the Best Offense Builds the Best Defense
SafeMind is CrowdStrike's closed-loop system where offensive agent Red Tempest and defensive agent Blue Solano co-evolve, yielding 70% improved accuracy and 6x faster detection creation.
blog · crowdstrike.com
CrowdStrike Agent Graph Shows How Agents Act on the Endpoint
CrowdStrike Agent Graph powers Falcon Guardian, connecting the AI layer to the execution layer to show not only what the agent was asked to do, but what it did.
blog · crowdstrike.com
CrowdStrike Expands Falcon Adversary OverWatch Threat Hunting Across AI and Google Cloud
Falcon Adversary OverWatch now includes 24/7 managed hunting for AI applications and autonomous agent activity, and deeper, telemetry-driven hunting for Google Cloud.
blog · crowdstrike.com
VAPOR PANDA Uses AI-Driven ARTEX to Scale a Global Espionage Campaign
CrowdStrike identified a VAPOR PANDA campaign using ARTEX, along with DeepSeek and another commercial LLM, to attack multiple high-value targets.
Back