Skip to main content
Mallory
CriticalExploited in the wildCISA KEVPublic exploitVulnerability

CVE-2025-22457: Unauthenticated Remote Code Execution in Ivanti Secure Access Gateways

Published Apr 3, 2025 · Updated 13 days ago
Trending this weekNot trendingOutside this week’s top 100
Mention activity · 30 days1 mention · peak 1 in a dayToday
CVSS9.8Critical · v3.1
EPSS100.0%100th percentile
CISA KEVYesAdded Apr 4, 2025
In the wildActivePer CISA KEV
Vendor
Ivanti
IdentifiersCVE-2025-22457CWE-121 · Stack-based Buffer Overflow

CVE-2025-22457 is a critical stack-based buffer overflow affecting Ivanti Connect Secure, Ivanti Policy Secure, Ivanti ZTA Gateways, and end-of-support Pulse Connect Secure 9.1x appliances. A remote unauthenticated attacker can exploit the overflow to execute arbitrary code. Affected releases include Connect Secure versions before 22.7R2.6 and ZTA Gateways versions before 22.8R2.2. The overflow was initially considered a low-risk denial-of-service issue because of restrictions on the input character space, but attackers developed a method to achieve code execution. Exploitation attributed to UNC5221, a suspected China-nexus espionage actor, was observed from at least mid-March 2025.

Do you run this?See affected versions in your environment and whether adversaries are exploiting this vulnerability against assets like yours.
Check your exposure
Analyst brief

Impact, mitigation & remediation

What it means and what to do now: patch path, interim mitigations, and the assume-compromise checklist.

Impact

What an attacker gets, and what they’ve been doing with it.

Successful exploitation can result in complete appliance compromise, affecting confidentiality, integrity, and availability. Observed attacks deployed the TRAILBLAZE in-memory dropper, BRUSHFIRE passive backdoor, and SPAWN malware, with attempts to tamper with integrity checking and logging. Compromised appliances can expose stored credentials, certificates, and plaintext secrets traversing the gateway, enabling espionage and lateral movement. Public proof-of-concept exploitation has also demonstrated web-server crashes.

Mitigation

If you can’t patch tonight, do this now.

Isolate potentially compromised appliances from both the internet and internal networks to limit further access and lateral movement. Preserve memory, filesystem snapshots, and logs where feasible before shutdown or recovery. Run both internal and external Ivanti Integrity Checker Tools, but supplement them with forensic investigation because attackers have attempted to tamper with integrity checking. Investigate web-process crashes, integrity-check anomalies, unusual client TLS certificates, outbound connections, and activity involving associated directory service accounts. If a corrective update cannot be applied, disconnect the appliance from the internet where possible and contact Ivanti support.

Remediation

Patch, then assume compromise.

Upgrade Ivanti Connect Secure to 22.7R2.6 or later and ZTA Gateways to 22.8R2.2 or later. Upgrade Policy Secure to a vendor-confirmed fixed release; the available information inconsistently identifies 22.7R1.4 and 22.7R1.5 as the fixed-version boundary. Decommission unsupported Pulse Connect Secure 9.1x appliances and migrate to a supported product. Investigate potential compromise even if patches have already been installed. Preserve forensic evidence before recovery. For confirmed compromise, perform vendor-directed factory reset or recovery and install corrected firmware, recognizing that a reset alone may not eliminate every threat. Rotate exposed administrative, local, service-account, directory, API, and license-server credentials; revoke affected authentication artifacts and reissue exposed certificates.

Public exploits

Exploits

6 exploits validated by Mallory's analysis. Hidden: 3 rejected by Mallory's analysis as not a working exploit.

View more in app →
ivanti_connect_secure_stack_overflow_rce_cve_2025_22457.rbThis repository contains a single Metasploit module targeting a stack-based buffer overflow vulnerability (CVE-2025-22457) in Ivanti Connect Secure (versions 22.7R2.5 and earlier) running on Linux. The exploit is unauthenticated and leverages a heap spray and brute-force approach to bypass ASLR and achieve remote code execution. The module interacts with the target's web interface over HTTPS (default port 443), first checking the product version via a GET request to '/dana-na/auth/url_admin/welcome.cgi'. If the target is vulnerable, the module attempts to exploit the buffer overflow by brute-forcing the base address of the 'libdsplibs' library, spraying memory, and triggering the overflow to execute a command shell payload (default: Meterpreter reverse shell). The exploit is operational and provides a remote shell if successful. The code is structured as a standard Metasploit module, with options for thread count, brute-force attempts, and target-specific parameters. No hardcoded IPs or domains are present; the module is designed to be used against user-specified targets.Disclosed Apr 3, 2025 · Framework: metasploit · ruby · networkWeaponizedAshwesker-CVE-2025-22457This repository contains a working exploit for CVE-2025-22457, a critical unauthenticated remote code execution vulnerability affecting Ivanti Connect Secure, Policy Secure, ZTA Gateways, and Pulse Connect Secure appliances. The exploit is implemented in Ruby (CVE-2025-22457.rb) and is accompanied by a detailed README.md explaining affected versions, exploitation steps, and mitigation advice. The exploit works by sending crafted HTTP/HTTPS requests to the target's web interface, fingerprinting the product version, and then performing a heap spray and buffer overflow attack against the /home/bin/web process, leveraging ROP gadgets in /home/lib/libdsplibs.so. The attacker must provide their own IP and port for the reverse shell payload, which is a standard bash reverse shell. The exploit brute-forces the base address of the target library unless provided directly, and can be tuned for the number of web child processes on the appliance. The code is operational and provides a real reverse shell if successful, but requires the attacker to have network access to the target's HTTPS port and to set up a listener. The exploit is not part of a framework and is a standalone PoC with a customizable payload. The endpoints targeted are the Ivanti web interface and specific internal files/processes on the appliance.Ashwesker · Disclosed Jul 13, 2025 · ruby · networkOperationalcve-2025-22457This repository is a self-contained lab simulation of a network-reachable stack-based buffer overflow, presented as a representative demo for CVE-2025-22457 rather than a real product exploit. The structure includes two pairs of C servers: a primary/demo pair (`meric_vulnerable_server.c` and `meric_patched_server.c`) and an older/simpler pair (`vulnerable_server.c` and `patched_server.c`). All servers implement a TCP service on port 9999; the vulnerable variants copy attacker-controlled network input into a 256-byte stack buffer without proper bounds checking (`strcpy`), while the patched variants add length checks or bounded copying. The main exploit capability is in `exploit.py`, which connects to `127.0.0.1:9999` and sends a classic overflow layout: 268 bytes of padding, a 4-byte EIP overwrite marker (`BBBB`), a short NOP sled, and placeholder INT3 bytes as demo shellcode. This indicates the exploit is intended to prove offset calculation and instruction-pointer control under a debugger, not to deliver a real post-exploitation payload. `crash_client.py` is a simpler companion tool that sends either a normal short string or a long repeated-character payload to demonstrate service stability versus crash behavior. The Makefile builds the `meric_*` server pair with insecure compilation settings suitable for a lab (`-fno-stack-protector`, `-no-pie`). The README files explain the educational purpose, compilation steps, and demo flow: start the vulnerable service, send benign traffic, send oversized traffic to trigger a segmentation fault, then compare behavior with the patched service. Overall, this is a genuine exploit/demo repository for a network attack vector, but it is clearly educational and local-lab oriented, with operational crash/EIP-control proof rather than weaponized remote code execution.benmevic · Disclosed Jul 9, 2026 · c · python · networkOperationalCVE-2025-22457This repository contains a proof-of-concept (PoC) exploit for CVE-2025-22457, a remote unauthenticated stack-based buffer overflow in Ivanti Connect Secure, Pulse Connect Secure, Policy Secure, and ZTA Gateway appliances (specifically version 22.7.2.3597). The exploit is implemented in Ruby (CVE-2025-22457.rb) and is designed to achieve remote code execution by sending crafted HTTPS requests to the target's web interface. The exploit brute-forces the base address of a shared library (libdsplibs.so) to construct a ROP chain, ultimately executing a user-supplied shell command (by default, a bash reverse shell). The attacker must provide the target's IP, port, and their own listener IP/port. The exploit includes options to tune the number of web child processes and the base address for more efficient exploitation. The ReadMe.md provides detailed usage instructions, expected results (reverse shell as 'nr' user), and technical background. The main attack vector is network-based, targeting the HTTPS service on the appliance. Notable endpoints include the version fingerprinting CGI and the use of internal binaries and libraries for exploitation.sfewer-r7 · Disclosed Apr 9, 2025 · ruby · markdown · networkPoCCVE-2025-22457This repository contains a proof-of-concept (PoC) exploit for CVE-2025-22457, a remote unauthenticated stack-based buffer overflow affecting Ivanti Connect Secure, Pulse Connect Secure, Ivanti Policy Secure, and ZTA Gateways (version 22.7r2.4). The main exploit is implemented in Ruby (CVE-2025-22457-main/CVE-2025-22457.rb) and leverages a network-based attack vector over HTTPS. The exploit works by brute-forcing the base address of a shared library (libdsplibs.so) and performing heap spraying to ensure reliable exploitation across multiple web child processes. Upon successful exploitation, the script executes a user-supplied payload (default: a bash reverse shell) on the target, granting remote shell access as the 'nr' user. The repository includes detailed usage instructions, debugging tips, and information about the target environment. No detection scripts or fake code are present; this is a functional PoC exploit.TRone-ux · Disclosed May 25, 2025 · ruby · networkPoCCVE-2025-22457This repository is a small lab-style stack buffer overflow simulation rather than a production exploit kit. It contains four files: a README with build guidance, exploit.py as the proof-of-concept client, vulnerable_server.c implementing the unsafe network service, and patched_server.c showing the remediated version. The vulnerable server listens on TCP port 9999, accepts up to 1024 bytes from a client, and passes the received data into vulnerable_function(), which copies it into a 256-byte stack buffer using strcpy without bounds checking. The exploit script connects to 127.0.0.1:9999 and sends a crafted payload consisting of 268 bytes of padding, a 4-byte EIP overwrite marker ('BBBB'), a short NOP sled, and 32 bytes of 0xCC INT3 instructions. This is designed for debugger-visible crash/control-flow validation, not post-exploitation. The patched server replaces strcpy with bounded strncpy and null termination, preserving the same network behavior while removing the overflow condition. Overall, the repository’s purpose is educational: demonstrate root cause, trigger conditions, and a basic exploit path for a stack-based overflow in a simplified embedded/firmware-like TCP service.melisakumral · Disclosed Jul 9, 2026 · python · c · network · localPoC
Exposure surface

Affected products & vendors

Products and vendors Mallory has correlated with this vulnerability. Open in Mallory to drill down to specific CPE configurations and version ranges.

VendorProductType
IvantiConnect Secureapplication
IvantiNeurons For Zero-Trust Accessapplication
IvantiPolicy Secureapplication
IvantiZero Trust Access Gatewayapplication
FAQ

Frequently asked questions

Is CVE-2025-22457 exploited in the wild?
Yes. As of Nov 4, 2026, Mallory has reports of exploitation in the wild, and CISA added it to KEV on Apr 4, 2025.
Is there a public exploit for CVE-2025-22457?
Yes. Mallory tracks 6 public exploits or proofs of concept for CVE-2025-22457, the earliest disclosed Apr 3, 2025.
Is there a patch for CVE-2025-22457?
Upgrade Ivanti Connect Secure to 22.7R2.6 or later and ZTA Gateways to 22.8R2.2 or later. Upgrade Policy Secure to a vendor-confirmed fixed release; the available information inconsistently identifies 22.7R1.4 and 22.7R1.5 as the fixed-version boundary.
What products are affected by CVE-2025-22457?
Ivanti Connect Secure, Ivanti Policy Secure, Ivanti Neurons For Zero-Trust Access, and Ivanti Zero Trust Access Gateway.
How severe is CVE-2025-22457?
CVE-2025-22457 has a CVSS v3.1 score of 9.8 (Critical) and an EPSS score of 100.0% (100th percentile) as of Nov 4, 2026.
Activity feed

Recent activity

74 sources tracked across advisories, community write-ups, and news. New activity surfaces here as Mallory finds it.

View more in app →
+42 more in the timeline. Sign up to see them
Do you run this?See affected versions in your environment and whether adversaries are exploiting this vulnerability against assets like yours.
Check your exposure
Mallory platform

The version that knows your environment.

This page is what’s public. Mallory adds the parts that aren’t: which of your assets are affected, which adversaries are exploiting it right now, and which detections to deploy tonight.
Exposure mappingQuery your assets running an affected version, and investigate the blast radius.
Threat actor evidence13Every report linking this CVE to a named adversary.
Associated malware28Malware families linked to this CVE, with evidence and IOCs.
Detection signatures2YARA, Sigma, Snort, and vendor rules, auto-deployed to your SIEM.
Vendor-by-vendor mappingCross-references every affected SKU, including bundled OEM variants.
Social activity22Community discussion across Reddit, Mastodon, and other social sources.