# GlassWorm VS Code Extension Supply Chain Attack - Open VSX Hijack

> GlassWorm is a supply chain attack campaign targeting VS Code extension registries — primarily the Open VSX Registry and to a lesser extent the official VS Code Marketplace — through malicious extensions that masquerade as popular developer tools. The campaign leverages VS Code's lack of sandboxing (extensions run with full user privileges), weak identity verification ('Verified' badge only proves domain ownership), and permissive naming policies (displayName can exactly replicate popular extensions) to distribute credential stealers, reverse shells, and data exfiltration payloads to software developers. Extensions activate on install and every VS Code startup, executing attacker-controlled code via eval() from remote C2 servers. The campaign is particularly dangerous because VS Code is used by 74%+ of developers (StackOverflow survey) and extensions have access to the full file system, SSH keys, Git credentials, source code repositories, and the ability to install additional malware. GlassWorm specifically targeted the Open VSX Registry which serves as the extension marketplace for VSCodium, GitPod, Eclipse Theia, and other open-source VS Code alternatives with less rigorous review processes.

- **Published:** 2026-02-02T21:40:00Z
- **Last reviewed:** 2026-02-02T21:40:00Z
- **Canonical:** https://intel.threadlinqs.com/threat/TL-2026-0024
- **Card image:** https://intel.threadlinqs.com/og/threat/threat-intel-glassworm-vs-code-supply-chain-tl-2026-0024.webp?v=9b65021a (Threadlinqs threat intel: GlassWorm VS Code Extension Hijack: Detection & IOCs (TL-2026-0024) — High severity, 28 IOCs, 13 detection rules.)
- **ID:** TL-2026-0024
- **Severity:** HIGH (CVSS 8.1)
- **Category:** SUPPLY_CHAIN
- **Status:** DORMANT
- **Actor:** GlassWorm
- **Detections:** 13 (1 shown free below) · **IOCs:** 28

## Description

GlassWorm represents an escalation in the IDE supply chain attack surface, combining the well-documented risks of package manager poisoning (npm/PyPI) with the uniquely dangerous permission model of VS Code extensions.

**VS Code Extension Security Model (or Lack Thereof):**

VS Code extensions run with the FULL privileges of the user that opened VS Code — there is NO sandbox. This means any extension can:
- Read, write, and delete any file accessible to the user
- Access SSH keys (~/.ssh/) and Git credentials (~/.gitconfig, credential helpers)
- Execute arbitrary shell commands via Node.js child_process
- Install additional software including ransomware, wipers, and backdoors
- Access and modify source code in all open repositories
- Read environment variables containing API keys and secrets
- Make network connections to any endpoint
- Access browser cookies and credential stores

This permission model makes VS Code extensions among the highest-impact supply chain attack vectors in the modern development ecosystem.

**The GlassWorm Campaign:**

The GlassWorm campaign deployed malicious extensions across the Open VSX Registry (open-vsx.org) and VS Code Marketplace through several attack techniques:

1. **Typosquatting:** Extensions published under names with subtle misspellings of popular extensions (e.g., 'pretier-vscode' instead of 'prettier-vscode', single-letter variations). When developers mistype extension names, they install the malicious variant.

2. **DisplayName Spoofing:** VS Code Marketplace and Open VSX allow arbitrary values in the 'displayName' field without uniqueness requirements. GlassWorm extensions used identical display names and icons to popular extensions (Prettier, Python, Jupyter), making them visually indistinguishable from legitimate versions.

3. **Verified Badge Abuse:** The VS Code Marketplace 'Verified' badge only confirms domain ownership — not publisher identity. GlassWorm operators registered cheap domains to obtain verified status, then changed their displayName to match popular publishers, creating convincing impersonation.

4. **Open VSX Targeting:** The Open VSX Registry (used by VSCodium, GitPod, Eclipse Theia, and other open-source VS Code alternatives) has less rigorous review processes than the Microsoft Marketplace. GlassWorm specifically targeted Open VSX as an easier ingress point for malicious extensions.

5. **Remote Code Execution via C2:** Malicious extensions included an activate() function that executed on installation and every VS Code startup. This function called out to attacker-controlled URLs (using hostname-specific subdomains for tracking: http://$HOSTNAME.attacker.com/vscode) and executed the response via eval(). This enabled:
   - Dynamic payload delivery (attacker changes server response to update malware)
   - Hostname-based targeting (deliver different payloads to different victims)
   - Anti-analysis evasion (benign response to researchers, malicious to targets)

6. **Credential Harvesting:** Extensions harvested:
   - SSH private keys from ~/.ssh/
   - Git credentials from credential helpers and .gitconfig
   - Environment variables containing AWS_ACCESS_KEY_ID, GITHUB_TOKEN, NPM_TOKEN
   - Browser cookies and stored passwords
   - VS Code settings containing API keys and service tokens

**Aqua Nautilus Research Context:**

Aqua Nautilus researchers demonstrated the viability of this attack vector by uploading a POC extension masquerading as Prettier (one of the top 10 most installed extensions with 27M+ installations). Within 48 hours, their POC received over 1,000 installations from active developers worldwide — demonstrating the scale of potential impact.

Aqua's research also identified existing suspicious extensions in the Marketplace, including 'API Generator Plugin' and 'code-tester', which contained code that contacted external C2 servers and executed responses via eval() every 30 seconds over unencrypted HTTP.

**Impact on Software Supply Chain:**

GlassWorm-compromised developers become vectors for downstream supply chain attacks:
- Stolen SSH keys enable direct commits to any repository the developer has access to
- Stolen NPM/PyPI tokens enable malicious package publication
- Modified source code on developer machines propagates to production
- Stolen cloud credentials (AWS, GCP, Azure) enable infrastructure compromise
- Lateral movement through Git organizations and CI/CD pipelines

This creates a cascading supply chain effect: compromise one developer → access all their repositories → inject malicious code → distribute to thousands of downstream consumers.

## MITRE ATT&CK

- T1195 Supply Chain Compromise
- T1199 Trusted Relationship
- T1059 Command and Scripting Interpreter
- T1204 User Execution
- T1546 Event Triggered Execution
- T1554 Compromise Host Software Binary
- T1036 Masquerading
- T1027 Obfuscated Files or Information
- T1070 Indicator Removal
- T1552 Unsecured Credentials
- T1528 Steal Application Access Token
- T1555 Credentials from Password Stores
- T1083 File and Directory Discovery
- T1082 System Information Discovery
- T1005 Data from Local System
- T1213 Data from Information Repositories
- T1071 Application Layer Protocol
- T1041 Exfiltration Over C2 Channel
- T1567 Exfiltration Over Web Service

## Free detection (Splunk SPL, Microsoft KQL)

**VS Code Extension Spawning Suspicious Process** (`TL-2026-0024-DET-001`)

Detects VS Code or extension host spawning suspicious child processes

- **Languages:** Splunk SPL, Microsoft KQL
- **Data sources:** sysmon, edr
- **Severity:** high
- **Confidence:** high

### Splunk SPL

```spl
index=sysmon sourcetype="XmlWinEventLog:Microsoft-Windows-Sysmon/Operational" EventCode=1
| where match(ParentImage, "(?i)(code\.exe|code-insiders\.exe|extensionHost)")
| where match(Image, "(?i)(cmd|powershell|wscript|cscript|mshta|certutil|bitsadmin|curl|wget)\.exe$")
   OR match(CommandLine, "(?i)(invoke-|downloadstring|downloadfile|webclient|http|base64)")
| stats count values(CommandLine) as commands by ParentImage, Image, Computer, User
| eval alert_severity="HIGH"
| eval alert_reason="VS Code spawned suspicious process - possible malicious extension"
| table Computer, User, ParentImage, Image, commands, count, alert_severity
```

### Microsoft KQL

```kql
// VS Code Extension Spawning Suspicious Process
// Detects VS Code or extension host spawning suspicious child processes
DeviceProcessEvents
| where Timestamp > ago(24h)
| where FileName has "(?i)(code\.exe|code-insiders\.exe|extensionHost)"
| where ProcessCommandLine contains "(?i)(invoke-|downloadstring|downloadfile|webclient|http|base64)"
| project Timestamp, DeviceName, AccountName, FileName, ProcessCommandLine, InitiatingProcessFileName
| summarize Count=count(), Commands=make_set(ProcessCommandLine) by DeviceName, AccountName, FileName
| where Count > 1
```

**Known false positives:**

- Legitimate extensions with terminal/shell functionality

**Tuning notes:** Whitelist known legitimate extension behaviors

### 12 more detection rules (Splunk SPL, Microsoft KQL, Sigma), unlock with TL Creds

- VS Code Extension Spawning Shell Process — GlassWorm Post-Install Execution (locked: unlock with TL Creds)
- VS Code Extension Accessing SSH Private Keys — Credential Theft (locked: unlock with TL Creds)
- VS Code Extension C2 Beacon — Periodic HTTP to Unknown Domain (locked: unlock with TL Creds)
- VS Code Extension Accessing SSH Keys on Linux — GlassWorm Credential Theft (locked: unlock with TL Creds)
- VS Code Extension Network Callback with Hostname in URL — C2 Tracking (locked: unlock with TL Creds)
- VS Code Extension Spawning Shell on Linux (locked: unlock with TL Creds)
- VS Code Extension Installing Additional Software — GlassWorm Payload Delivery (locked: unlock with TL Creds)
- VS Code Extension Reading Git Credentials — Developer Token Theft (locked: unlock with TL Creds)
- VS Code Extension Using eval() with Network-Fetched Data — Remote Code Execution (locked: unlock with TL Creds)
- VS Code Extension Network Callback (locked: unlock with TL Creds)
- VS Code Extension Modification After Install (locked: unlock with TL Creds)
- Recently Installed VS Code Extension Audit (locked: unlock with TL Creds)

Sign up free to unlock the detection rules: https://threadlinqs.com/?modal=signup

## Sources

- [Aqua Nautilus — Can You Trust Your VSCode Extensions?](https://www.aquasec.com/blog/can-you-trust-your-vscode-extensions/)
- [VS Code Marketplace — Extension Trust Guidelines](https://code.visualstudio.com/docs/editor/extension-marketplace#_can-i-trust-extensions-from-the-marketplace)
- [Open VSX Registry](https://open-vsx.org/)
- [StackOverflow Developer Survey 2022 — VS Code Dominance](https://survey.stackoverflow.co/2022/#section-most-popular-technologies-integrated-development-environment)
- [MITRE ATT&CK — Masquerading (T1036)](https://attack.mitre.org/techniques/T1036/)
- [MITRE CAPEC — Typosquatting (CAPEC-630)](https://capec.mitre.org/data/definitions/630.html)
- [VS Code Extension API — Activation Events](https://code.visualstudio.com/api/references/activation-events)

## Full data

The other 12 detection rules (Splunk SPL / Microsoft KQL / Sigma) unlock with TL Creds; a free account includes daily creds: https://threadlinqs.com/?modal=signup
Indicator values unlock with TL Creds, at a lower price on Red and Purple; AI agents can query rules and IOCs through the Threadlinqs MCP server (Purple): https://intel.threadlinqs.com/mcp

Canonical: https://intel.threadlinqs.com/threat/TL-2026-0024
