mirror of
				https://github.com/HackTricks-wiki/hacktricks.git
				synced 2025-10-10 18:36:50 +00:00 
			
		
		
		
	Add content from: You name it, VMware elevates it (CVE-2025-41244)
This commit is contained in:
		
							parent
							
								
									5b50426b39
								
							
						
					
					
						commit
						504b1597dc
					
				| @ -108,6 +108,7 @@ | ||||
| - [Checklist - Linux Privilege Escalation](linux-hardening/linux-privilege-escalation-checklist.md) | ||||
| - [Linux Privilege Escalation](linux-hardening/privilege-escalation/README.md) | ||||
|   - [Android Rooting Frameworks Manager Auth Bypass Syscall Hook](linux-hardening/privilege-escalation/android-rooting-frameworks-manager-auth-bypass-syscall-hook.md) | ||||
|   - [Vmware Tools Service Discovery Untrusted Search Path Cve 2025 41244](linux-hardening/privilege-escalation/vmware-tools-service-discovery-untrusted-search-path-cve-2025-41244.md) | ||||
|   - [Arbitrary File Write to Root](linux-hardening/privilege-escalation/write-to-root.md) | ||||
|   - [Cisco - vmanage](linux-hardening/privilege-escalation/cisco-vmanage.md) | ||||
|   - [Containerd (ctr) Privilege Escalation](linux-hardening/privilege-escalation/containerd-ctr-privilege-escalation.md) | ||||
|  | ||||
| @ -1,6 +1,6 @@ | ||||
| # Mutation Testing for Solidity with Slither (slither-mutate) | ||||
| 
 | ||||
| {{#include ../../../banners/hacktricks-training.md}} | ||||
| {{#include ../../banners/hacktricks-training.md}} | ||||
| 
 | ||||
| Mutation testing "tests your tests" by systematically introducing small changes (mutants) into your Solidity code and re-running your test suite. If a test fails, the mutant is killed. If the tests still pass, the mutant survives, revealing a blind spot in your test suite that line/branch coverage cannot detect. | ||||
| 
 | ||||
| @ -123,4 +123,4 @@ Guidance: Treat survivors that affect value transfers, accounting, or access con | ||||
| - [Arkis DeFi Prime Brokerage Security Review (Appendix C)](https://github.com/trailofbits/publications/blob/master/reviews/2024-12-arkis-defi-prime-brokerage-securityreview.pdf) | ||||
| - [Slither (GitHub)](https://github.com/crytic/slither) | ||||
| 
 | ||||
| {{#include ../../../banners/hacktricks-training.md}} | ||||
| {{#include ../../banners/hacktricks-training.md}} | ||||
|  | ||||
| @ -1723,6 +1723,16 @@ Android rooting frameworks commonly hook a syscall to expose privileged kernel f | ||||
| android-rooting-frameworks-manager-auth-bypass-syscall-hook.md | ||||
| {{#endref}} | ||||
| 
 | ||||
| ## VMware Tools service discovery LPE (CWE-426) via regex-based exec (CVE-2025-41244) | ||||
| 
 | ||||
| Regex-driven service discovery in VMware Tools/Aria Operations can extract a binary path from process command lines and execute it with -v under a privileged context. Permissive patterns (e.g., using \S) may match attacker-staged listeners in writable locations (e.g., /tmp/httpd), leading to execution as root (CWE-426 Untrusted Search Path). | ||||
| 
 | ||||
| Learn more and see a generalized pattern applicable to other discovery/monitoring stacks here: | ||||
| 
 | ||||
| {{#ref}} | ||||
| vmware-tools-service-discovery-untrusted-search-path-cve-2025-41244.md | ||||
| {{#endref}} | ||||
| 
 | ||||
| ## Kernel Security Protections | ||||
| 
 | ||||
| - [https://github.com/a13xp0p0v/kconfig-hardened-check](https://github.com/a13xp0p0v/kconfig-hardened-check) | ||||
| @ -1774,4 +1784,6 @@ android-rooting-frameworks-manager-auth-bypass-syscall-hook.md | ||||
| - [GNU Bash Manual – BASH_ENV (non-interactive startup file)](https://www.gnu.org/software/bash/manual/bash.html#index-BASH_005fENV) | ||||
| - [0xdf – HTB Environment (sudo env_keep BASH_ENV → root)](https://0xdf.gitlab.io/2025/09/06/htb-environment.html) | ||||
| 
 | ||||
| - [NVISO – You name it, VMware elevates it (CVE-2025-41244)](https://blog.nviso.eu/2025/09/29/you-name-it-vmware-elevates-it-cve-2025-41244/) | ||||
| 
 | ||||
| {{#include ../../banners/hacktricks-training.md}} | ||||
|  | ||||
| @ -0,0 +1,153 @@ | ||||
| # VMware Tools service discovery LPE (CWE-426) via regex-based binary discovery (CVE-2025-41244) | ||||
| 
 | ||||
| {{#include ../../banners/hacktricks-training.md}} | ||||
| 
 | ||||
| This technique abuses regex-driven service discovery pipelines that parse running process command lines to infer service versions and then execute a candidate binary with a "version" flag. When permissive patterns accept untrusted, attacker-controlled paths (e.g., /tmp/httpd), the privileged collector executes an arbitrary binary from an untrusted location, yielding local privilege escalation. NVISO documented this in VMware Tools/Aria Operations Service Discovery as CVE-2025-41244. | ||||
| 
 | ||||
| - Impact: Local privilege escalation to root (or to the privileged discovery account) | ||||
| - Root cause: Untrusted Search Path (CWE-426) + permissive regex matching of process command lines | ||||
| - Affected: open-vm-tools/VMware Tools on Linux (credential-less discovery), VMware Aria Operations SDMP (credential-based discovery via Tools/proxy) | ||||
| 
 | ||||
| ## How VMware service discovery works (high level) | ||||
| 
 | ||||
| - Credential-based (legacy): Aria executes discovery scripts inside the guest via VMware Tools using configured privileged credentials. | ||||
| - Credential-less (modern): Discovery logic runs within VMware Tools, already privileged in the guest. | ||||
| 
 | ||||
| Both modes ultimately run shell logic that scans processes with listening sockets, extracts a matching command path via a regex, and executes the first argv token with a version flag. | ||||
| 
 | ||||
| ## Root cause and vulnerable pattern (open-vm-tools) | ||||
| 
 | ||||
| In open-vm-tools, the serviceDiscovery plugin script get-versions.sh matches candidate binaries using broad regular expressions and executes the first token without any trusted-path validation: | ||||
| 
 | ||||
| ```bash | ||||
| get_version() { | ||||
|   PATTERN=$1 | ||||
|   VERSION_OPTION=$2 | ||||
|   for p in $space_separated_pids | ||||
|   do | ||||
|     COMMAND=$(get_command_line $p | grep -Eo "$PATTERN") | ||||
|     [ ! -z "$COMMAND" ] && echo VERSIONSTART "$p" "$("${COMMAND%%[[:space:]]*}" $VERSION_OPTION 2>&1)" VERSIONEND | ||||
|   done | ||||
| } | ||||
| ``` | ||||
| 
 | ||||
| It is invoked with permissive patterns containing \S (non-whitespace) that will happily match non-system paths in user-writable locations: | ||||
| 
 | ||||
| ```bash | ||||
| get_version "/\S+/(httpd-prefork|httpd|httpd2-prefork)($|\s)" -v | ||||
| get_version "/usr/(bin|sbin)/apache\S*" -v | ||||
| get_version "/\S+/mysqld($|\s)" -V | ||||
| get_version "\.?/\S*nginx($|\s)" -v | ||||
| get_version "/\S+/srm/bin/vmware-dr($|\s)" --version | ||||
| get_version "/\S+/dataserver($|\s)" -v | ||||
| ``` | ||||
| 
 | ||||
| - Extraction uses grep -Eo and takes the first token: ${COMMAND%%[[:space:]]*} | ||||
| - No whitelist/allowlist of trusted system paths; any discovered listener with a matching name is executed with -v/--version | ||||
| 
 | ||||
| This creates an untrusted search path execution primitive: arbitrary binaries located in world-writable directories (e.g., /tmp/httpd) get executed by a privileged component. | ||||
| 
 | ||||
| ## Exploitation (both credential-less and credential-based modes) | ||||
| 
 | ||||
| Preconditions | ||||
| - You can run an unprivileged process that opens a listening socket on the guest. | ||||
| - The discovery job is enabled and runs periodically (historically ~5 minutes). | ||||
| 
 | ||||
| Steps | ||||
| 1) Stage a binary in a path matching one of the permissive regexes, e.g. /tmp/httpd or ./nginx | ||||
| 2) Run it as a low-privileged user and ensure it opens any listening socket | ||||
| 3) Wait for the discovery cycle; the privileged collector will automatically execute: /tmp/httpd -v (or similar), running your program as root | ||||
| 
 | ||||
| Minimal demo (using NVISO’s approach) | ||||
| ```bash | ||||
| # Build any small helper that: | ||||
| #  - default mode: opens a dummy TCP listener | ||||
| #  - when called with -v/--version: performs the privileged action (e.g., connect to an abstract UNIX socket and spawn /bin/sh -i) | ||||
| # Example staging and trigger | ||||
| cp your_helper /tmp/httpd | ||||
| chmod +x /tmp/httpd | ||||
| /tmp/httpd          # run as low-priv user and wait for the cycle | ||||
| # After the next cycle, expect a root shell or your privileged action | ||||
| ``` | ||||
| 
 | ||||
| Typical process lineage | ||||
| - Credential-based: /usr/bin/vmtoolsd -> /bin/sh /tmp/VMware-SDMP-Scripts-.../script_...sh -> /tmp/httpd -v -> /bin/sh -i | ||||
| - Credential-less: /bin/sh .../get-versions.sh -> /tmp/httpd -v -> /bin/sh -i | ||||
| 
 | ||||
| Artifacts (credential-based) | ||||
| Recovered SDMP wrapper scripts under /tmp/VMware-SDMP-Scripts-{UUID}/ may show direct execution of the rogue path: | ||||
| ```bash | ||||
| /tmp/httpd -v >"/tmp/VMware-SDMP-Scripts-{UUID}/script_-{ID}_0.stdout" 2>"/tmp/VMware-SDMP-Scripts-{UUID}/script_-{ID}_0.stderr" | ||||
| ``` | ||||
| 
 | ||||
| ## Generalizing the technique: regex-driven discovery abuse (portable pattern) | ||||
| 
 | ||||
| Many agents and monitoring suites implement version/service discovery by: | ||||
| - Enumerating processes with listening sockets | ||||
| - Grepping argv/command lines with permissive regexes (e.g., patterns containing \S) | ||||
| - Executing the matched path with a benign flag like -v, --version, -V, -h | ||||
| 
 | ||||
| If the regex accepts untrusted paths and the path is executed from a privileged context, you get CWE-426 Untrusted Search Path execution. | ||||
| 
 | ||||
| Abuse recipe | ||||
| - Name your binary like common daemons that the regex is likely to match: httpd, nginx, mysqld, dataserver | ||||
| - Place it in a writable directory: /tmp/httpd, ./nginx | ||||
| - Ensure it matches the regex and opens any port to be enumerated | ||||
| - Wait for the scheduled collector; you get an automatic privileged invocation of <path> -v | ||||
| 
 | ||||
| Masquerading note: This aligns with MITRE ATT&CK T1036.005 (Match Legitimate Name or Location) to increase match probability and stealth. | ||||
| 
 | ||||
| Reusable privileged I/O relay trick | ||||
| - Build your helper so that on privileged invocation (-v/--version) it connects to a known rendezvous (e.g., a Linux abstract UNIX socket like @cve) and bridges stdio to /bin/sh -i. This avoids on-disk artifacts and works across many environments where the same binary is re-invoked with a flag. | ||||
| 
 | ||||
| ## Detection and DFIR guidance | ||||
| 
 | ||||
| Hunting queries | ||||
| - Uncommon children of vmtoolsd or get-versions.sh such as /tmp/httpd, ./nginx, /tmp/mysqld | ||||
| - Any execution of non-system absolute paths by discovery scripts (look for spaces in ${COMMAND%%...} expansions) | ||||
| - ps -ef --forest to visualize ancestry trees: vmtoolsd -> get-versions.sh -> <non-system path> | ||||
| 
 | ||||
| On Aria SDMP (credential-based) | ||||
| - Inspect /tmp/VMware-SDMP-Scripts-{UUID}/ for transient scripts and stdout/stderr artifacts showing execution of attacker paths | ||||
| 
 | ||||
| Policy/telemetry | ||||
| - Alert when privileged collectors execute from non-system prefixes: ^/(tmp|home|var/tmp|dev/shm)/ | ||||
| - File integrity monitoring on get-versions.sh and VMware Tools plugins | ||||
| 
 | ||||
| ## Mitigations | ||||
| 
 | ||||
| - Patch: Apply Broadcom/VMware updates for CVE-2025-41244 (Tools and Aria Operations SDMP) | ||||
| - Disable or restrict credential-less discovery where feasible | ||||
| - Validate trusted paths: restrict execution to allowlisted directories (/usr/sbin, /usr/bin, /sbin, /bin) and only exact known binaries | ||||
| - Avoid permissive regexes with \S; prefer anchored, explicit absolute paths and exact command names | ||||
| - Drop privileges for discovery helpers where possible; sandbox (seccomp/AppArmor) to reduce impact | ||||
| - Monitor for and alert on vmtoolsd/get-versions.sh executing non-system paths | ||||
| 
 | ||||
| ## Notes for defenders and implementers | ||||
| 
 | ||||
| Safer matching and execution pattern | ||||
| ```bash | ||||
| # Bad: permissive regex and blind exec | ||||
| COMMAND=$(get_command_line "$pid" | grep -Eo "/\\S+/nginx(\$|\\s)") | ||||
| [ -n "$COMMAND" ] && "${COMMAND%%[[:space:]]*}" -v | ||||
| 
 | ||||
| # Good: strict allowlist + path checks | ||||
| candidate=$(get_command_line "$pid" | awk '{print $1}') | ||||
| case "$candidate" in | ||||
|   /usr/sbin/nginx|/usr/sbin/httpd|/usr/sbin/apache2) | ||||
|       "$candidate" -v 2>&1 ;; | ||||
|   *) | ||||
|       : # ignore non-allowlisted paths | ||||
|       ;; | ||||
| esac | ||||
| ``` | ||||
| 
 | ||||
| ## References | ||||
| 
 | ||||
| - [NVISO – You name it, VMware elevates it (CVE-2025-41244)](https://blog.nviso.eu/2025/09/29/you-name-it-vmware-elevates-it-cve-2025-41244/) | ||||
| - [Broadcom advisory for CVE-2025-41244](https://support.broadcom.com/web/ecx/support-content-notification/-/external/content/SecurityAdvisories/0/36149) | ||||
| - [open-vm-tools – serviceDiscovery/get-versions.sh (stable-13.0.0)](https://github.com/vmware/open-vm-tools/blob/stable-13.0.0/open-vm-tools/services/plugins/serviceDiscovery/get-versions.sh) | ||||
| - [MITRE ATT&CK T1036.005 – Match Legitimate Name or Location](https://attack.mitre.org/techniques/T1036/005/) | ||||
| - [CWE-426: Untrusted Search Path](https://cwe.mitre.org/data/definitions/426.html) | ||||
| 
 | ||||
| {{#include ../../banners/hacktricks-training.md}} | ||||
| @ -20,6 +20,15 @@ msf> auxiliary/scanner/vmware/vmware_http_login | ||||
| If you find valid credentials, you can use more metasploit scanner modules to obtain information. | ||||
| 
 | ||||
| 
 | ||||
| 
 | ||||
| ### See also | ||||
| 
 | ||||
| Linux LPE via VMware Tools service discovery (CWE-426 / CVE-2025-41244): | ||||
| 
 | ||||
| {{#ref}} | ||||
| ../../linux-hardening/privilege-escalation/vmware-tools-service-discovery-untrusted-search-path-cve-2025-41244.md | ||||
| {{#endref}} | ||||
| 
 | ||||
| {{#include ../../banners/hacktricks-training.md}} | ||||
| 
 | ||||
| 
 | ||||
|  | ||||
		Loading…
	
	
			
			x
			
			
		
	
		Reference in New Issue
	
	Block a user