A buffer overflow vulnerability exists in TP-Link TL-WR940N V3 and V4 routers running firmware version 3.16.9 and earlier. The flaw is present in the handling of the dnsserver1 and dnsserver2 parameters at the /userRpm/Wan6to4TunnelCfgRpm.htm endpoint. An authenticated attacker can exploit this vulnerability to execute arbitrary code on the device as the root user. The vulnerability is due to improper bounds checking on user-supplied input to these parameters, leading to memory corruption and control flow hijacking.
Mallory correlates every CVE against your assets, your vendors, and active adversary campaigns. Know which vulnerabilities matter for you, not just which ones are loud.
What it means. What to do now. Patch path, mitigations, and the assume-compromise checklist.
What an attacker gets, and what they’ve been doing with it.
If you can’t patch tonight, do this now.
Patch, then assume compromise.
1 valid exploit after Mallory filtered fakes, detection scripts, and README-only repos.
This repository is a small standalone TypeScript proof-of-concept exploit for CVE-2024-54887 targeting TP-Link TL-WR940N routers (README states v3/v4). The repository contains one substantive code file, cve-2024-54887.ts, plus standard Node/TypeScript project metadata (package.json, package-lock.json, tsconfig.json) and a README. The exploit flow is straightforward: it parses CLI arguments for target IP and optional credentials, defaults to admin/admin, computes an MD5 hash of the password, and sends a GET request to /userRpm/LoginRpm.htm?Save=Save with an Authorization cookie containing Base64(username:md5(password)). It then parses the returned HTML to extract the randomized authenticated session path from a redirect pattern like /<token>/userRpm/. Using that session URL, it sends a second GET request to Wan6to4TunnelCfgRpm.htm with a maliciously oversized dnsserver1 parameter. The payload is operational rather than a mere detector: it includes hardcoded MIPS shellcode and several hardcoded big-endian addresses used as a ROP chain (derived from a fixed libc base). The shellcode appears intended to execute /bin/sh in a bind-shell configuration, and both the README and script indicate the operator should check TCP port 4444 after exploitation. Because the payload and gadget addresses are embedded and not dynamically adapted, this is best classified as OPERATIONAL rather than weaponized. Notable fingerprintable targets are the router’s HTTP admin endpoints /userRpm/LoginRpm.htm and /<session-token>/userRpm/Wan6to4TunnelCfgRpm.htm, along with the expected post-exploitation bind shell on port 4444. Overall, the repository’s purpose is to provide a runnable ts-node rewrite of a public Python PoC for authenticated remote code execution against the vulnerable router web interface.
Products and vendors Mallory has correlated with this vulnerability. Open in Mallory to drill down to specific CPE configurations and version ranges.
Vendor-confirmed product mapping. Mallory continuously reconciles this list against your asset inventory.
No public activity tracked yet. Mallory keeps watching.
No public activity observed for this vulnerability.
Query your assets running an affected version, and investigate the blast radius.
Every observed campaign linking this CVE to a named adversary.
Malware families riding this exploit, with evidence and IOCs.
YARA, Sigma, Snort, and vendor rules, auto-deployed to your SIEM.
Cross-references every affected SKU, including bundled OEM variants.
Community discussion across Reddit, Mastodon, and other social sources.