Use Node JS to scan files on your server with ClamAV's clamscan/clamdscan binary or via TCP to a remote server or local UNIX Domain socket. This is especially useful for scanning uploaded files provided by un-trusted sources.
If you are using a version prior to 1.2.0, please upgrade! There was a security vulnerability in previous versions that can cause false negative in some edge cases. Specific details on how the attack could be implemented will not be disclosed here. Please update to 1.2.0 or greater ASAP. No breaking changes are included, only the security patch.
All older versions in NPM have been deprecated.
If you are migrating from v0.8.5 or less to v1.0.0 or greater, please read the release notes as there are some breaking changes (but also some awesome new features!).
- Dependencies
- How to Install
- License Info
- Getting Started
- Basic Usage Example
- API
- Contribute
- Resources used to help develop this module
You will need to install ClamAV's clamscan binary and/or have clamdscan daemon running on your server. On linux, it's quite simple.
Fedora-based distros:
sudo yum install clamavDebian-based distros:
sudo apt-get install clamav clamav-daemonFor OS X, you can install clamav with brew:
sudo brew install clamavYou will need access to either:
- A local UNIX Domain socket for a local instance of
clamd
- Follow instructions in To use local binary method of scanning.
- Socket file is usually:
/var/run/clamd.scan/clamd.sock - Make sure
clamdis running on your local server
- A local/remote
clamddaemon
- Must know the port the daemon is running on
- If running on remote server, you must have the IP address/domain name
- If running on remote server, it's firewall must have the appropriate TCP port(s) open
- Make sure
clamdis running on your local/remote server
NOTE: This module is not intended to work on a Windows server. This would be a welcome addition if someone wants to add that feature (I may get around to it one day but have no urgent need for this).
npm install clamscanLicensed under the MIT License:
All of the values listed in the example below represent the default values for their respective configuration item.
You can simply do this:
const NodeClam = require('clamscan');
const ClamScan = new NodeClam().init();And, you'll be good to go.
BUT: If you want more control, you can specify all sorts of options.
const NodeClam = require('clamscan');
const ClamScan = new NodeClam().init({
removeInfected: false, // If true, removes infected files
quarantineInfected: false, // False: Don't quarantine, Path: Moves files to this place.
scanLog: null, // Path to a writeable log file to write scan results into
debugMode: false, // Whether or not to log info/debug/error msgs to the console
fileList: null, // path to file containing list of files to scan (for scanFiles method)
scanRecursively: true, // If true, deep scan folders recursively
clamscan: {
path: '/usr/bin/clamscan', // Path to clamscan binary on your server
db: null, // Path to a custom virus definition database
scanArchives: true, // If true, scan archives (ex. zip, rar, tar, dmg, iso, etc...)
active: true // If true, this module will consider using the clamscan binary
},
clamdscan: {
socket: false, // Socket file for connecting via TCP
host: false, // IP of host to connect to TCP interface
port: false, // Port of host to use when connecting via TCP interface
timeout: 60000, // Timeout for scanning files
localFallback: true, // Use local preferred binary to scan if socket/tcp fails
path: '/usr/bin/clamdscan', // Path to the clamdscan binary on your server
configFile: null, // Specify config file if it's in an unusual place
multiscan: true, // Scan using all available cores! Yay!
reloadDb: false, // If true, will re-load the DB on every call (slow)
active: true, // If true, this module will consider using the clamdscan binary
bypassTest: false, // Check to see if socket is available when applicable
tls: false, // Use plaintext TCP to connect to clamd
},
preference: 'clamdscan' // If clamdscan is found and active, it will be used by default
});Here is a non-default values example (to help you get an idea of what proper-looking values could be):
const NodeClam = require('clamscan');
const ClamScan = new NodeClam().init({
removeInfected: true, // Removes files if they are infected
quarantineInfected: '~/infected/', // Move file here. removeInfected must be FALSE, though.
scanLog: '/var/log/node-clam', // You're a detail-oriented security professional.
debugMode: true, // This will put some debug info in your js console
fileList: '/home/webuser/scanFiles.txt', // path to file containing list of files to scan
scanRecursively: false, // Choosing false here will save some CPU cycles
clamscan: {
path: '/usr/bin/clam', // I dunno, maybe your clamscan is just call "clam"
scanArchives: false, // Choosing false here will save some CPU cycles
db: '/usr/bin/better_clam_db', // Path to a custom virus definition database
active: false // you don't want to use this at all because it's evil
},
clamdscan: {
socket: '/var/run/clamd.scan/clamd.sock', // This is pretty typical
host: '127.0.0.1', // If you want to connect locally but not through socket
port: 12345, // Because, why not
timeout: 300000, // 5 minutes
localFallback: false, // Do no fail over to binary-method of scanning
path: '/bin/clamdscan', // Special path to the clamdscan binary on your server
configFile: '/etc/clamd.d/daemon.conf', // A fairly typical config location
multiscan: false, // You hate speed and multi-threaded awesome-sauce
reloadDb: true, // You want your scans to run slow like with clamscan
active: false, // you don't want to use this at all because it's evil
bypassTest: true, // Don't check to see if socket is available. You should probably never set this to true.
tls: true, // Connect to clamd over TLS
},
preference: 'clamscan' // If clamscan is found and active, it will be used by default
});NOTE: If a valid port is provided but no host value is provided, the clamscan will assume 'localhost' for host.
As of version v1.0.0, this module supports communication with a local or remote ClamAV daemon through Unix Domain sockets or a TCP host/port combo. If you supply both in your configuration object, the UNIX Domain socket option will be used. The module will not not fallback to using the alternative Host/Port method. If you wish to connect via Host/Port and not a Socket, please either omit the socket property in the config object or use socket: null.
If you specify a valid clamscan/clamdscan binary in your config and you set clamdscan.localFallback: true in your config, this module will fallback to the traditional way this module has worked--using a binary directly/locally.
Also, there are some caveats to using the socket/tcp based approach:
-
The following configuration items are not honored (unless the module falls back to binary method):
removeInfected- remote clamd service config will dictate thisquarantineInfected- remote clamd service config will dictate thisscanLog- remote clamd service config will dictate thisfileList- this simply won't be availableclamscan.db- only available on fallbackclamscan.scanArchives- only available on fallbackclamscan.path- only available on fallbackclamdscan.configFile- only available on fallbackclamdscan.path- only available on fallback
For the sake of brevity, all the examples in the API section will be shortened to just the relevant parts related specifically to that example. In those examples, we'll assume you already have an instance of the clamscan object. Since initializing the module returns a promise, you'll have to resolve that promise to get an instance of the clamscan object.
Below is the full example of how you could get that instance and run some methods:
const NodeClam = require('clamscan');
const ClamScan = new NodeClam().init(options);
// Get instance by resolving ClamScan promise object
ClamScan.then(async clamscan => {
try {
// You can re-use the `clamscan` object as many times as you want
const version = await clamscan.getVersion();
console.log(`ClamAV Version: ${version}`);
const {isInfected, file, viruses} = await clamscan.isInfected('/some/file.zip');
if (isInfected) console.log(`${file} is infected with ${viruses}!`);
} catch (err) {
// Handle any errors raised by the code in the try block
}
}).catch(err => {
// Handle errors that may have occurred during initialization
});If you're writing your code within an async function, getting an instance can be one less step:
const NodeClam = require('clamscan');
async some_function() {
try {
// Get instance by resolving ClamScan promise object
const clamscan = await new NodeClam().init(options);
const {goodFiles, badFiles} = await clamscan.scanDir('/foo/bar');
} catch (err) {
// Handle any errors raised by the code in the try block
}
}
some_function();Complete/functional examples for various use-cases can be found in the examples folder.
This method allows you to determine the version of ClamAV you are interfacing with. It supports a callback and Promise API. If no callback is supplied, a Promise will be returned.
-
callback(function) (optional) Will be called when the scan is complete. It receives 2 parameters:err(object or null) A standard javascript Error object (null if no error)version(string) The version of the clamav server you're interfacing with
-
Promise
- Promise resolution returns:
version(string) The version of the clamav server you're interfacing with
- Promise resolution returns:
clamscan.getVersion((err, version) => {
if (err) return console.error(err);
console.log(`ClamAV Version: ${version}`);
});clamscan.getVersion().then(version => {
console.log(`ClamAV Version: ${version}`);
}).catch(err => {
console.error(err);
});This method allows you to scan a single file. It supports a callback and Promise API. If no callback is supplied, a Promise will be returned. This method will likely be the most common use-case for this module.
.scan_file
-
filePath(string) Represents a path to the file to be scanned. -
callback(function) (optional) Will be called when the scan is complete. It takes 3 parameters:err(object or null) A standard javascript Error object (null if no error)file(string) The originalfilePathpassed into theisInfectedmethod.isInfected(boolean) True: File is infected; False: File is clean. NULL: Unable to scan.viruses(array) An array of any viruses found in the scanned file.
-
Promise
-
Promise resolution returns:
result(object):file(string) The originalfilePathpassed into theisInfectedmethod.isInfected(boolean) True: File is infected; False: File is clean. NULL: Unable to scan.viruses(array) An array of any viruses found in the scanned file.
-
clamscan.isInfected('/a/picture/for_example.jpg', (err, file, isInfected, viruses) => {
if (err) return console.error(err);
if (isInfected) {
console.log(`${file} is infected with ${viruses.join(', ')}.`);
}
});clamscan.isInfected('/a/picture/for_example.jpg').then(result => {
const {file, isInfected, viruses} = result;
if (isInfected) console.log(`${file} is infected with ${viruses.join(', ')}.`);
}).then(err => {
console.error(err);
})const {file, isInfected, viruses} = await clamscan.isInfected('/a/picture/for_example.jpg');Allows you to scan an entire directory for infected files. This obeys your recursive option even for clamdscan which does not have a native way to turn this feature off. If you have multiple paths, send them in an array to scanFiles.
TL;DR: For maximum speed, don't supply a fileCallback.
If you choose to supply a fileCallback, the scan will run a little bit slower (depending on number of files to be scanned) for clamdscan. If you are using clamscan, while it will work, I'd highly advise you to NOT pass a fileCallback... it will run incredibly slow.
The goodFiles parameter of the endCallback callback in this method will only contain the directory that was scanned in all but the following scenarios:
- A
fileCallbackcallback is provided, andscanRecursivelyis set to true. - The scanner is set to
clamdscanandscanRecursivelyis set to false. - The scanned directory contains 1 or more viruses. In this case, the
goodFilesarray will be empty.
There will, however, be a total count of the good files which is calculated by determining the total number of files scanned and subtracting the number of bad files from that count. We simply can't provide a list of all good files due to the potential large memory usage implications of scanning a directory with, for example, millions of files.
-
dirPath(string) (required) Full path to the directory to scan. -
endCallback(function) (optional) Will be called when the entire directory has been completely scanned. This callback takes 3 parameters:err(object) A standard javascript Error object (null if no error)goodFiles(array) An empty array if path is infected. An array containing the directory name that was passed in if clean.badFiles(array) List of the full paths to all files that are infected.viruses(array) List of all the viruses found (feature request: associate to the bad files).numGoodFiles(number) Number of files that were found to be clean.
-
fileCallback(function) (optional) Will be called after each file in the directory has been scanned. This is useful for keeping track of the progress of the scan. This callback takes 3 parameters:err(object or null) A standard Javascript Error object (null if no error)file(string) Path to the file that just got scanned.isInfected(boolean) True: File is infected; False: File is clean. NULL: Unable to scan file.
-
Promise
-
Promise resolution returns:
result(object):path(string) The originaldir_pathpassed into thescanDirmethod.isInfected(boolean) True: File is infected; False: File is clean. NULL: Unable to scan.goodFiles(array) An empty array if path is infected. An array containing the directory name that was passed in if clean.badFiles(array) List of the full paths to all files that are infected.viruses(array) List of all the viruses found (feature request: associate to the bad files).numGoodFiles(number) Number of files that were found to be clean.
-
clamscan.scanDir('/some/path/to/scan', (err, goodFiles, badFiles, viruses, numGoodFiles) {
if (err) return console.error(err);
if (badFiles.length > 0) {
console.log(`${path} was infected. The offending files (${badFiles.join (', ')}) have been quarantined.`);
console.log(`Viruses Found: ${viruses.join(', ')}`);
} else {
console.log(`${goodFiles[0]} looks good! ${numGoodFiles} file scanned and no problems found!.`);
}
});clamscan.scanDir('/some/path/to/scan').then(results => {
const { path, isInfected, goodFiles, badFiles, viruses, numGoodFiles } = results;
//...
}).catch(err => {
return console.error(err);
});const { path, isInfected, goodFiles, badFiles, viruses, numGoodFiles } = await clamscan.scanDir('/some/path/to/scan');This allows you to scan many files that might be in different directories or maybe only certain files of a single directory. This is essentially a wrapper for isInfected that simplifies the process of scanning many files or directories.
-
files(array) (optional) A list of strings representing full paths to files you want scanned. If not supplied, the module will check for afileListconfig option. If neither is found, the method will throw an error. -
endCallback(function) (optional) Will be called when the entire list of files has been completely scanned. This callback takes 3 parameters:err(object or null) A standard JavaScript Error object (null if no error)goodFiles(array) List of the full paths to all files that are clean.badFiles(array) List of the full paths to all files that are infected.
-
fileCallback(function) (optional) Will be called after each file in the list has been scanned. This is useful for keeping track of the progress of the scan. This callback takes 3 parameters:err(object or null) A standard JavaScript Error object (null if no error)file(string) Path to the file that just got scanned.isInfected(boolean) True: File is infected; False: File is clean. NULL: Unable to scan file.
-
Promise
-
Promise resolution returns:
result(object):goodFiles(array) List of the full paths to all files that are clean.badFiles(array) List of the full paths to all files that are infected.errors(object) Per-file errors keyed by the filename in which the error happened. (ex.{'foo.txt': Error})viruses(array) List of all the viruses found (feature request: associate to the bad files).
-
const scan_status = { good: 0, bad: 0 };
const files = [
'/path/to/file/1.jpg',
'/path/to/file/2.mov',
'/path/to/file/3.rb'
];
clamscan.scanFiles(files, (err, goodFiles, badFiles, viruses) => {
if (err) return console.error(err);
if (badFiles.length > 0) {
console.log({
msg: `${goodFiles.length} files were OK. ${badFiles.length} were infected!`,
badFiles,
goodFiles,
viruses,
});
} else {
res.send({msg: "Everything looks good! No problems here!."});
}
}, (err, file, isInfected, viruses) => {
;(isInfected ? scan_status.bad++ : scan_status.good++);
console.log(`${file} is ${(isInfected ? `infected with ${viruses}` : 'ok')}.`);
console.log('Scan Status: ', `${(scan_status.bad + scan_status.good)}/${files.length}`);
});Note: There is currently no way to get per-file notifications with the Promise API.
clamscan.scanFiles(files).then(results => {
const { goodFiles, badFiles, errors, viruses } = results;
// ...
}).catch(err => {
console.error(err);
})const { goodFiles, badFiles, errors, viruses } = await clamscan.scanFiles(files);If this modules is configured with a valid path to a file containing a newline-delimited list of files, it will use the list in that file when scanning if the first paramter passed is falsy.
Files List Document:
/some/path/to/file.zip
/some/other/path/to/file.exe
/one/more/file/to/scan.rbScript:
const ClamScan = new NodeClam().init({
fileList: '/path/to/fileList.txt'
});
ClamScan.then(async clamscan => {
// Supply nothing to first parameter to use `fileList`
const { goodFiles, badFiles, errors, viruses } = await clamscan.scanFiles();
});This method allows you to scan a binary stream. NOTE: This method will only work if you've configured the module to allow the use of a TCP or UNIX Domain socket. In other words, this will not work if you only have access to a local ClamAV binary.
-
stream(stream) A readable stream object -
callback(function) (optional) Will be called after the stream has been scanned (or attempted to be scanned):err(object or null) A standard JavaScript Error object (null if no error)isInfected(boolean) True: Stream is infected; False: Stream is clean. NULL: Unable to scan file.
-
Promise
-
Promise resolution returns:
result(object):file(string) NULL as no file path can be provided with the streamisInfected(boolean) True: File is infected; False: File is clean. NULL: Unable to scan.viruses(array) An array of any viruses found in the scanned file.
-
Callback Example:
const NodeClam = require('clamscan');
// You'll need to specify your socket or TCP connection info
const clamscan = new NodeClam().init({
clamdscan: {
socket: '/var/run/clamd.scan/clamd.sock',
host: '127.0.0.1',
port: 3310,
}
});
const Readable = require('stream').Readable;
const rs = Readable();
rs.push('foooooo');
rs.push('barrrrr');
rs.push(null);
clamscan.scanStream(stream, (err, { isInfected. viruses }) => {
if (err) return console.error(err);
if (isInfected) return console.log('Stream is infected! Booo!', viruses);
console.log('Stream is not infected! Yay!');
});Promise Example:
clamscan.scanStream(stream).then(({isInfected}) => {
if (isInfected) return console.log("Stream is infected! Booo!");
console.log("Stream is not infected! Yay!");
}).catch(err => {
console.error(err);
};Promise Example:
const { isInfected, viruses } = await clamscan.scanStream(stream);The passthrough method returns a PassthroughStream object which allows you pipe a ReadbleStream through it and on to another output. In the case of this module's passthrough implementation, it's actually forking the data to also go to ClamAV via TCP or Domain Sockets. Each data chunk is only passed on to the output if that chunk was successfully sent to and received by ClamAV. The PassthroughStream object returned from this method has a special event that is emitted when ClamAV finishes scanning the streamed data so that you can decide if there's anything you need to do with the final output destination (ex. delete a file or S3 object).
In typical, non-passthrough setups, a file is uploaded to the local filesytem and then subsequently scanned. With that setup, you have to wait for the upload to complete and then wait again for the scan to complete. Using this module's passthrough method, you could theoretically speed up user uploads intended to be scanned by up to 2x because the files are simultaneously scanned and written to any WriteableStream output (examples: filesystem, S3, gzip, etc...).
As for these theoretical gains, your mileage my vary and I'd love to hear feedback on this to see where things can still be improved.
Please note that this method is different than all the others in that it returns a PassthroughStream object and does not support a Promise or Callback API. This makes sense once you see the example below (a practical working example can be found in the examples directory of this module):
const NodeClam = require('clamscan');
// You'll need to specify your socket or TCP connection info
const clamscan = new NodeClam().init({
clamdscan: {
socket: '/var/run/clamd.scan/clamd.sock',
host: '127.0.0.1',
port: 3310,
}
});
// For example's sake, we're using the Axios module
const axios = require('Axios');
// Get a readable stream for a URL request
const input = axios.get(some_url);
// Create a writable stream to a local file
const output = fs.createWriteStream(some_local_file);
// Get instance of this module's PassthroughStream object
const av = clamscan.passthrough();
// Send output of Axios stream to ClamAV.
// Send output of Axios to `some_local_file` if ClamAV receives data successfully
input.pipe(av).pipe(output);
// What happens when scan is completed
av.on('scan-complete', result => {
const { isInfected, viruses } = result;
// Do stuff if you want
});
// What happens when data has been fully written to `output`
output.on('finish', () => {
// Do stuff if you want
});
// NOTE: no errors (or other events) are being handled in this example but standard errors will be emitted according to NodeJS's Stream specificationsGot a missing feature you'd like to use? Found a bug? Go ahead and fork this repo, build the feature and issue a pull request.
- https://stuffivelearned.org/doku.php?id=apps:clamav:general:remoteclamdscan
- http://cpansearch.perl.org/src/JMEHNLE/ClamAV-Client-0.11/lib/ClamAV/Client.pm
- https://github.com/yongtang/clamav.js
- https://nodejs.org/dist/latest-v10.x/docs/api/stream.html
- https://manpages.debian.org/jessie/clamav-daemon/clamd.8.en.html
![dependabot[bot] avatar](https://avatars.githubusercontent.com/in/29110?v=4 "dependabot[bot]")
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent