# China's Webworm APT Weaponizes Discord and Microsoft Graph to Target EU Government Infrastructure


A sophisticated Chinese-linked advanced persistent threat (APT) group known as Webworm has launched a sustained campaign against European government entities, leveraging legitimate cloud communication services and tunneling infrastructure to evade detection. The campaign highlights a troubling trend: adversaries increasingly weaponize widely-trusted platforms—Discord for command-and-control (C2) operations and Microsoft Graph API for reconnaissance—while using encrypted tunneling protocols to mask malicious traffic from network defenders.


Security researchers tracking the operation have identified a multi-stage attack chain that combines social engineering, credential theft, and living-off-the-land techniques to establish persistent access within government networks across the European Union. The use of consumer-grade communication tools and legitimate APIs represents a significant shift in APT tradecraft, enabling attackers to blend malicious traffic with billions of legitimate connections.


## The Threat: Webworm's EU Government Campaign


Webworm, a state-sponsored APT group with confirmed ties to Chinese intelligence operations, has been actively targeting European government agencies since at least early 2026. The campaign represents a strategic escalation in cyber espionage operations against NATO-aligned governments, with initial victims identified in diplomatic, defense, and foreign ministry networks across multiple EU member states.


Key campaign characteristics:

  • Timeline: Active operations confirmed from January 2026 to present
  • Targets: EU government agencies across at least 5 member states
  • Primary objectives: Intelligence gathering, credential harvesting, lateral movement capability
  • Infection vector: Spear-phishing emails with weaponized Office documents containing embedded malware

  • The group's focus on EU governments aligns with Beijing's broader strategic interest in European geopolitical intelligence, trade policy coordination, and NATO military posture. Unlike financially-motivated cybercriminals, Webworm operates with the resources and patience of a nation-state, maintaining persistence even after partial compromise discovery.


    ## The Attack Chain: How It Works


    The Webworm campaign follows a methodical progression designed to establish long-term access while remaining undetected:


    ### Stage 1: Initial Compromise

    Victims receive spear-phishing emails tailored to government personnel with administrative privileges. Attachments appear to be legitimate documents—DOCX files themed around EU policy, defense reports, or trade negotiations. When opened, malicious macros execute and download a first-stage payload.


    ### Stage 2: Reconnaissance and Credential Harvesting

    The initial implant performs network reconnaissance, identifying the victim's organizational structure, security posture, and accessible cloud services. Attackers specifically enumerate:

  • Active Directory enumeration: Discovering domain structure, user accounts, and group memberships
  • Cloud service discovery: Identifying Microsoft 365, OneDrive, SharePoint, and other cloud services
  • VPN and proxy infrastructure: Mapping network segmentation and external connectivity

  • ### Stage 3: Microsoft Graph API Exploitation

    Rather than attempting to breach Microsoft 365 directly, Webworm abuses the Microsoft Graph API—a legitimate Microsoft service that allows authorized applications to query organizational data. By obtaining a single compromised credential, attackers can:

  • Query organizational structure without raising flags
  • Enumerate mailboxes to identify high-value targets
  • Access calendar information to understand meeting patterns and operational activity
  • Retrieve SharePoint metadata to identify classified or sensitive document repositories

  • This approach is particularly insidious because Graph API queries generate minimal security alerts—they appear as normal business activity from inside the network.


    ### Stage 4: C2 Communication via Discord

    Once foothold is established, Webworm establishes persistent command-and-control infrastructure using Discord, the popular voice and text communication platform. By creating seemingly innocuous Discord servers and channels, attackers:

  • Blend C2 traffic with millions of legitimate Discord connections
  • Avoid signature-based detection (Discord appears in allowlists across most corporate networks)
  • Enable two-way communication for command execution and data exfiltration
  • Maintain operational security by rotating servers and accounts

  • Discord's free hosting, global CDN, and encryption make it an ideal C2platform for threat actors. Traffic analysis becomes nearly impossible without deep inspection, and blocking Discord entirely disrupts legitimate employee communication.


    ## The Infrastructure: SOCKS Proxies and SoftEther VPN


    To mask the origin of outbound connections and evade geolocation-based detection, Webworm deploys SOCKS (Socket Secure) proxy infrastructure. SOCKS proxies function as intermediaries, forwarding traffic between a client and remote server while obscuring the true source IP address.


    Specifically, the campaign leverages SoftEther VPN, an open-source tunnel utility that:

  • Creates encrypted tunnels through restrictive firewalls and proxies
  • Operates over HTTPS (port 443), making it nearly indistinguishable from legitimate HTTPS traffic
  • Supports SOCKS proxy protocols for flexible redirection
  • Is widely available and difficult to distinguish from legitimate remote access tools

  • By chaining multiple SOCKS proxies and SoftEther tunnels, Webworm operators effectively hide their command servers behind layers of intermediary infrastructure. Each proxy adds latency but provides deniability—the true origin becomes impossible to determine without extensive threat intelligence.


    ## Technical Details and Attack Evolution


    Researchers analyzing the Webworm malware samples report several notable technical characteristics:


    | Aspect | Detail |

    |--------|--------|

    | Malware Family | Custom PowerShell-based implant with modular architecture |

    | Persistence Mechanism | Scheduled Task registration with randomized names |

    | Living-off-the-Land | Extensive use of Windows Management Instrumentation (WMI) and PowerShell |

    | Anti-Analysis | Runtime code obfuscation, API hashing, and antivirus evasion techniques |

    | Lateral Movement | Pass-the-hash attacks using compromised credential material |


    The modular design allows Webworm operators to add new capabilities—data exfiltration, privilege escalation exploits, network scanning—without redeploying the entire malware framework.


    ## Why This Works: The Perfect Storm


    The effectiveness of the Webworm campaign stems from a convergence of security gaps:


    Trusted services: Discord and Microsoft Graph API are legitimate tools used billions of times daily. Network defenders cannot simply block them without disrupting operations.


    API abuse: Microsoft Graph lacks granular controls to detect reconnaissance queries at scale. An attacker querying calendars and organizational structure appears identical to a legitimate administrator.


    Encryption at scale: VPN traffic, proxy connections, and encrypted messengers mask attacker infrastructure from network monitoring.


    Human factors: Spear-phishing remains effective against government personnel, who may not recognize socially-engineered emails themed around genuine policy concerns.


    ## Implications for Organizations


    This campaign exposes critical vulnerabilities in how governments and enterprises protect sensitive information:


  • Cloud services are not network perimeter: Attackers don't need to breach firewalls anymore; they need one valid credential to abuse cloud APIs
  • Legitimate tools enable crime: Services like Discord and SoftEther VPN have no malicious purpose, yet become force multipliers for APTs
  • Reconnaissance is silent: API-based information gathering generates minimal security alerts
  • Living-off-the-land hides intent: Using native OS tools (PowerShell, WMI) defeats signature-based detection

  • ## Recommendations for Defenders


    Organizations targeted by or at risk from similar operations should implement:


    ### Immediate Actions

  • Enforce multi-factor authentication (MFA) on all cloud services and administrative accounts
  • Review Graph API permissions and remove unnecessary delegated access scopes
  • Monitor Graph API activity for bulk enumeration queries (large batches of organizational data requests)
  • Implement conditional access policies to flag logins from unusual locations or impossible travel scenarios

  • ### Mid-Term Measures

  • Deploy cloud access security brokers (CASBs) to monitor and control Microsoft 365 usage
  • Conduct credential audit: Force password resets for all administrative accounts
  • Segment network access: Implement zero-trust architecture to limit lateral movement
  • Block or monitor Discord and similar communication platforms at the network level (though this carries operational risk)

  • ### Long-Term Strategy

  • Assume breach mentality: Assume adversaries have compromised at least one credential; design defenses accordingly
  • API security program: Establish processes for auditing cloud API usage and permissions
  • Threat intelligence integration: Subscribe to intelligence feeds specific to Webworm and Chinese APT operations
  • Incident response drills: Test organizational ability to detect and respond to this specific attack chain

  • ## Looking Beyond This Campaign


    The Webworm operation reflects a broader trend in APT evolution: nation-states are optimizing for *operational ease* and *evading detection* rather than technical complexity. By abusing legitimate services and tools, Webworm gains several advantages: plausible deniability, minimal maintenance, global scalability, and natural obfuscation.


    This pattern will likely accelerate. Expect more APT campaigns to exploit Discord, Telegram, Microsoft Teams, and similar platforms for C2 operations. Expect more API abuse targeting Azure, AWS, and Google Cloud. The security industry's focus on blocking malware signatures has left a gaping hole: legitimate business tools weaponized by state actors.


    ---


    ## HackWire Analysis


    The Webworm campaign represents a watershed moment in APT tradecraft—not because it introduces novel attack techniques, but because it demonstrates how comprehensively existing defenses can be bypassed using unmodified, commercially available software.


    The decision to use Discord for command-and-control is particularly telling. It signals that Webworm operators have concluded traditional C2 infrastructure (dedicated servers, bulletproof hosting, custom protocols) is too risky. A Discord server, by contrast, can be created in seconds, costs nothing, and appears identical to millions of legitimate business communications. Even if discovered, investigators face the challenge of attributing the communication to a specific operator rather than a compromised employee account.


    The abuse of Microsoft Graph API reveals a critical blind spot in cloud security. Most organizations monitor for *failed authentication attempts* and *permission escalations*, but not for suspicious API query patterns. An attacker who has obtained a legitimate credential can freely enumerate the entire organizational structure, identify sensitive data repositories, and extract metadata—all while appearing as an authorized user. Defenders are essentially blind to reconnaissance conducted via legitimate APIs.


    The timing is also significant: as governments and enterprises accelerate cloud migration, the surface area for Graph API abuse expands exponentially. Every organization moving to Microsoft 365 inherits this risk unless it implements explicit controls to detect and prevent reconnaissance queries.


    For defenders, the uncomfortable truth is that blocking Discord or restricting Microsoft Graph is operationally infeasible for most organizations. Employees use Discord for project coordination. Administrators use Graph APIs for legitimate provisioning tasks. Blanket bans cripple productivity. Instead, defenders must adopt a more sophisticated approach: assume these tools *will* be compromised, and focus on detecting *misuse patterns* rather than blocking the tools themselves.


    — HackWire Editorial


    ---


    ## Related Coverage


  • Read more in our [Breaches](https://www.hackwire.news/category/breaches) coverage
  • Cross-reference with [Vulnerabilities](https://www.hackwire.news/category/vulnerabilities) and [Advanced Threats](https://www.hackwire.news/category/advanced-threats)
  • Stay current via the [HackWire homepage](https://www.hackwire.news/)