CVE-2025-29635 is a command injection flaw in the /goform/set_prohibiting endpoint of D-Link DIR-823X router firmware that lets an attacker run arbitrary OS commands on the device, the router is already end-of-life with no patch coming, and Akamai has observed it actively exploited in the wild by a Mirai botnet variant since March 2026.
Vulnerability summary
CVSS base score
N/A
No NVD score published as of
EPSS score (FIRST)
87.9%
Probability of exploitation in the wild in the next 30 days
CISA KEV catalog
Yes
Listed as known exploited
Published
Vulnerability ID: CVE-2025-29635 · Source: NVD
Affected components
| Component | CPE | Version range |
|---|---|---|
| D-Link DIR-823X (firmware) | cpe:2.3:o:dlink:dir-823x_firmware:240126:*:*:*:*:*:*:* | Firmware 240126 and 240802; device is end-of-life/end-of-service since 2024-11-15, no fixed version available |
What does this mean for affected devices?
CVE-2025-29635 (CWE-77, improper neutralization of special elements used in a command) is a command injection vulnerability in D-Link DIR-823X router firmware versions 240126 and 240802. NVD's description states it "allows an authorized attacker to execute arbitrary commands on remote devices by sending a POST request to /goform/set_prohibiting via the corresponding function, triggering remote command execution" (NVD).
The only CVSS metric on record is a CVSS v3.1 score of 7.2 (high), vector CVSS:3.1/AV:N/AC:L/PR:H/UI:N/S:U/C:H/I:H/A:H, attributed to CISA's NVD-authorized data publisher (ADP) feed rather than to NVD's own analysts, and NVD has not published an independent Primary score. Because that figure is a Secondary (ADP) metric rather than an NVD-confirmed Primary score, it is not presented here as an NVD-confirmed CVSS score. The PR:H (high privileges required) component of that vector should be read against the real-world exploitation pattern described below, which does not document a credential requirement for the observed attacks (NVD).
Akamai's security research team reports the root cause in technical detail: "the value of the macaddr function is copied into the command variable by the snprintf function, after which the system function is executed" — i.e., attacker-controlled input reaches a shell command unsanitized. Akamai's honeypot network first observed active exploitation in March 2026, a year after the CVE's initial March 2025 disclosure, delivering a Mirai botnet variant the researchers named "tuxnokill." The malware supports multiple CPU architectures, uses XOR-encoded strings, has hard-coded command-and-control infrastructure at 64.89.161[.]130:44300, and conducts standard DDoS attacks including TCP SYN floods, UDP DNS floods, and HTTP floods (Akamai, CVE-2025-29635: Mirai Campaign Targets D-Link Devices).
D-Link's own support advisory confirms the DIR-823X, across all hardware revisions, reached End of Life/End of Service as of 2024-11-15, states that D-Link "will be unable to resolve device or firmware issues since all development and customer support has ceased," and "strongly recommends that this product be retired." No patched firmware exists or is scheduled (D-Link Security Advisement, SAP10469).
CISA added CVE-2025-29635 to the Known Exploited Vulnerabilities (KEV) catalog on 2026-04-24, with a required-action due date of 2026-05-08, and its KEV entry explicitly notes the affected product "could be end-of-life (EoL) and/or end-of-service (EoS)" and that "users should discontinue product utilization" (CISA KEV).
As of 2026-10-29, FIRST scores the probability of exploitation at 0.87944 (99.763th percentile), among the highest EPSS scores observed and consistent with ongoing botnet scanning (FIRST EPSS).
Because there is no fix and the device is formally end-of-life, any DIR-823X still in service should be retired and replaced rather than patched; if continued use is unavoidable in the short term, isolate it from the internet-facing interface, disable remote management, and monitor for the documented Mirai indicators of compromise.
Evidence
- Command injection in D-Link DIR-823X firmware 240126 and 240802: attacker sends a crafted POST request to /goform/set_prohibiting to execute arbitrary commands. — NVD
- The only CVSS metric (3.1, 7.2, vector including PR:H) is a CISA-ADP Secondary score; NVD has not published its own Primary score. — NVD
- Root cause: unsanitized macaddr value copied via snprintf into a command string executed by system(); actively exploited since March 2026 by a Mirai variant ('tuxnokill') with hard-coded C2 at 64.89.161.130:44300, conducting SYN/UDP/HTTP flood DDoS. — Akamai, CVE-2025-29635: Mirai Campaign Targets D-Link Devices
- DIR-823X (all hardware revisions) reached End of Life/End of Service on 2024-11-15; D-Link will not resolve device or firmware issues and recommends retiring the product. — D-Link Security Advisement, SAP10469
- Added to the CISA Known Exploited Vulnerabilities catalog on 2026-04-24, required-action due date 2026-05-08; KEV entry flags the product as possibly EoL/EoS and recommends discontinuing use. — CISA KEV
- EPSS score is 0.87944 (99.763th percentile) as of 2026-10-29. — FIRST EPSS

Related advisories
Sources
Frequently asked questions
Does exploiting this vulnerability actually require high privileges, as the published CVSS vector suggests?
The only CVSS metric on record — a Secondary score from CISA's NVD-authorized data publisher feed, not an NVD-confirmed Primary score — uses a vector with PR:H (high privileges required). Per Akamai's technical analysis of the real-world attacks, the observed exploitation pattern does not describe any credential requirement, so that PR:H component should be read cautiously against the documented attacks.
What name did researchers give to the botnet exploiting this flaw, and how long after disclosure did exploitation begin?
Per Akamai, its honeypot network first observed active exploitation in March 2026 — about a year after the CVE's initial March 2025 disclosure — delivering a Mirai botnet variant the researchers named "tuxnokill."
What does the malware do once it compromises a vulnerable router?
Per Akamai, the "tuxnokill" payload supports multiple CPU architectures, uses XOR-encoded strings, communicates with hard-coded command-and-control infrastructure at 64.89.161[.]130:44300, and conducts standard DDoS attacks including TCP SYN floods, UDP DNS floods and HTTP floods.
Will D-Link ever release a patch for this vulnerability?
No — per D-Link's own security advisement SAP10469, the DIR-823X (all hardware revisions) reached End of Life/End of Service on 2024-11-15, and D-Link states it will be unable to resolve device or firmware issues since development and support have ceased, recommending the product be retired instead.
How does FIRST's exploitation-probability score for this CVE compare to a typical vulnerability?
Per FIRST EPSS, as of 2026-10-29 CVE-2025-29635 scores 0.87944 (99.763rd percentile) — among the highest scores observed in this batch, consistent with the sustained Mirai botnet scanning Akamai documented.
Questions about this page?