By Hamza Faisal

Answering an ordinary Telegram voice call could have been enough to let an attacker run code on an Android phone, according to security researchers who disclosed the flaw on October 29. The catch: the company behind the app says the underlying bug was already fixed weeks earlier, and that no one has shown a working attack chain.

Security startup NebuSec says the flaw sat in Telegram for Android's call-handling code and could be chained with two other bugs to take full control of a victim's phone, including its messages, microphone and camera. The attacker needed only a regular Telegram account, and the victim only had to pick up the call.

A four-year-old bug in the call code

The problem lived in tgcalls, Telegram's shared native calling library, in the code that parses a signalling message called MediaState. During a call, the two phones exchange short messages reporting the state of the media session, such as whether the microphone is muted, whether the camera is on, and how video should be rotated.

According to NebuSec, the parser checked that one value, the video rotation, was present, then read a second value, the video state, without checking that it existed at all. That mishandling of missing fields let the code access memory it should not have touched, and NebuSec says it turned that mistake into remote code execution on ARM64 Android phones after the victim answered a crafted call.

The flaw affects Telegram for Android versions 8.6.2 through 12.10.3, released between March 2022 and September 2026, which means the vulnerable code went unnoticed for about four years. The Hacker News, reviewing Telegram's published source code for version 12.10.3, found a flaw matching NebuSec's description.

Telegram disputes the exploit claim

Telegram fixed the underlying logic error in version 12.10.4, released in late September. NebuSec reported the flaw on September 26, after that version had shipped, and says it received no bug bounty because Telegram did not believe the flaw could be turned into remote code execution. The flaw has no CVE identifier.

On October 30, Telegram pushed back further. Spokesperson Remi Vaughn said the company had received a report about the flaw in August and had already patched it before NebuSec reported it, adding that neither report demonstrated a working remote code execution chain. NebuSec says its Code Security product, which runs on OpenAI's GPT-5.4 model, found the flaw, and its demonstration video shows a victim phone leaving the Telegram interface after answering a call, though the clip does not show which app or Android versions were running.

NebuSec says it has no evidence the flaw was used in real attacks, and no evidence it directly affects Telegram's iOS or desktop apps or group calls.

What users should do

The fix is simple: make sure Telegram for Android is on version 12.10.4 or later. The version number appears at the bottom of the app's Settings page. NebuSec also notes that setting Telegram's call privacy option to "Nobody" blocks the attack, as does simply declining an unexpected call.

Separately, Trend Micro's Zero Day Initiative published an October 15 advisory for a different Telegram issue, CVE-2026-111105, a flaw in the rlottie library used to render animated stickers. That advisory is unrelated to the call-handling flaw, but it is a reminder that even everyday interactions in a messaging app, a call, a sticker, a file, can carry risk when the code behind them mishandles unexpected input.