Win-DoS
A comprehensive framework implementing the "Win-DoS Epidemic" research, which discovered 5 critical vulnerabilities affecting Windows systems through novel attack techniques.
Created by SafeBreach Labs (Joint work of Or Yair and Shahak Morag). For the full technical analysis check out the blog post
Overview
This framework implements research that discovered 5 critical vulnerabilities affecting Windows systems:
Win-DoS Vulnerabilities (4)
- LDAP Referral DoS (CVE-2025-32724) - Memory exhaustion via massive LDAP referral lists affecting Domain Controllers
- NetLogon RPC DoS #1 (CVE-2025-26673) - Memory exhaustion in
NetrServerReqChallengefunction affecting Domain Controllers - NetLogon RPC DoS #2 (CVE-2025-49716) - Memory exhaustion in
DsrAddressToSiteNamesWfunction affecting Domain Controllers - Spoolsv RPC DoS (CVE-2025-49722) - Memory exhaustion in
RpcEnumPrintersfunction affecting all Windows endpoints
Win-DDoS Vulnerability (1)
- LDAP Referral DDoS (CVE-2025-32724) - Leveraging Domain Controllers as DDoS botnet participants without authentication
Setup
-
Clone the repository:
git clone <repository-url> cd Win-DoS -
Install dependencies for specific modules:
- For LDAP attacks:
cd ldap_attacks && pip install -r requirements.txt - For RPC attacks:
cd rpc_attacks && pip install -r requirements.txt
- For LDAP attacks:
Module Documentation
LDAP Attacks
Implements the LDAP referral attacks we discovered. Features:
- DoS Mode: Memory exhaustion via massive LDAP referral lists (500,000+ URLs)
- DDoS Mode: Leveraging Domain Controllers as botnet participants
RPC Attacks
Implements the RPC memory exhaustion attacks we discovered. Features:
- TorpeDoS technique: Pre-bind thousands of clients without waiting for bind acks, pre-sign packets if needed, and flood victims with many RPC calls at once
Authors - Or Yair & Shahak Morag
| Or Yair | Shahak Morag | |
|---|---|---|
| Or Yair | Shahak Morag | |
| @oryair1999 | @shahakmo |
Description
A set of tools and exploits to cause DoS for remote Windows Server & Windows 11 machines
Languages
Python
100%