Article
Ghostwriter: Hybrid Influence and Espionage Operations in Eastern Europe
Ghostwriter (UNC1151, TA445, UAC-0057) profile: a Belarus-aligned group active since 2016 mixing disinformation, credential theft and espionage.
Quick links
Press Enter to search, Esc to close
Significant cyber incident
Ghostwriter, which CERT Polska tracks as UNC1151, has run phishing, malware and disinformation campaigns against people and bodies in Poland since at least mid 2021.
In short
Ghostwriter, which CERT Polska tracks as UNC1151, has run phishing, malware and disinformation campaigns against people and bodies in Poland since at least mid 2021. CERT Polska's reports for 2024, 2025 and June 2026 describe stolen mailbox logins, an exploited webmail flaw, election related disinformation and, from March 2026, phishing aimed at Gmail users. Targets are politicians, officials, journalists and researchers.
On 19 July 2022 CERT Polska, the Polish national response team run by the research institute NASK, wrote that the group it calls UNC1151/Ghostwriter had for over a year been attacking the mailboxes of Polish citizens. It counted nearly 100 domains set up in the previous year for attacks on Polish speakers. In one case the attacks reached the mailboxes of members of parliament.
CERT Polska’s annual report for 2024 says the group kept up this activity. For most of the year it ran an e-mail campaign that used account suspension and password changes as the excuse. It also sent malware inside files that looked like invitations to industry events. In June 2024 CERT Polska worked on a case at a Polish national sports body. It found the intruder had probably used a flaw in one of the systems and then resold the access to others. Data about Polish athletes was stolen, and at the end of the Paris Olympics it was published on a Telegram channel linked to the group, together with disinformation about doping.
In 2025 the group repeatedly used a flaw in the Roundcube webmail program (CVE-2024-42009) to steal logins. CERT Polska reported the first such campaign on 5 June 2025. In March 2025 CSIRT NASK also saw the group running disinformation around the presidential election. On 12 June 2026 CERT Polska reported that since March 2026 the group had been phishing Gmail users, including their second login code.
CERT Polska says published work links UNC1151 to the Belarusian government, and that other sources point to Russian intelligence. See the attribution note.
CERT Polska says UNC1151/Ghostwriter has for over a year been attacking the mailboxes of Polish citizens.
CERT Polska counts nearly 100 domains set up in the previous year for attacks on Polish speakers.
The group runs a large e-mail campaign using account suspension and password change messages.
The group sends malware in files that look like invitations to industry events; infected systems download Cobalt Strike.
CERT Polska speaks with a Polish national sports body about a possible malware infection.
The body's website is hacked and stolen athlete data is published on a Telegram channel linked to the group.
CSIRT NASK sees UNC1151 activity tied to the upcoming presidential elections.
CERT Polska reports a spear phishing campaign that exploits the Roundcube flaw CVE-2024-42009 to steal logins.
The group repeatedly exploits the same flaw; APT28 also uses it against Polish entities.
The group starts phishing campaigns aimed at Gmail users.
CERT Polska publishes its analysis of the Gmail campaign and says the second login factor is also phished.
The e-mails pretend to come from a mail provider. In 2024 and 2025 CERT Polska describes messages about account suspension and password changes, aimed at users of Polish providers such as Interia, Onet and Wirtualna Polska. From March 2026 the same kind of message targeted Gmail users. CERT Polska says the group sometimes does not know whose mailbox it is writing to. It also says the campaigns run regularly, from Monday to Friday, over many weeks. The technique matches Phishing (T1566).
A fake login page collects the address and password. In the Gmail campaign the page also asks for the second login code, and the attackers then try to log in automatically. In the June 2025 campaign no fake page was needed. Opening the e-mail was enough to exploit the Roundcube flaw CVE-2024-42009. The exploit installed a Service Worker, a browser feature that runs in the background. It copied every genuine login to a server the attackers controlled.
CERT Polska says the attackers then read the mailbox, download the address book and sometimes send more phishing from the account. Taken mailboxes are searched for sensitive documents and used to take over linked social media accounts. The group reused real messages from taken accounts and attached the exploit to them, so new targets saw a mail they could trust. Using stolen logins matches Valid Accounts (T1078).
In malware attacks the group used files that looked like invitations to industry events, mostly CHM help files in 2025. Infected systems downloaded Cobalt Strike. In 2025 it also used a fake CAPTCHA, known as ClickFix, on agriculture related websites it had gained access to.
The group also publishes material. In March 2025 it used the image of a Ukrainian foundation and of election candidates, promoted it through YouTube and TikTok advertising and posted it in OLX auction descriptions. CERT Polska says the aims included deepening polarisation and inciting hostility toward the Ukrainian minority in Poland.
CERT Polska, 8 April 2026 (annual report for 2025): says UNC1151 is a cluster it attributes to APT activity, that published work links the group to the Belarusian government, and that other sources also point to connections with Russian intelligence services.
CERT Polska, 5 June 2025: attributes the Roundcube campaign to a UNC1151 cluster with high confidence, based on technical indicators.
Sources differ on which state stands behind the group: Belarus in the first view, Russian intelligence in the other. CERT Polska reports both. The Polish government statements on this group were not opened for this page.
CERT Polska gives no total number of victims. It says the 2024 e-mail activity was on a large scale and that the 2025 e-mail campaigns targeted hundreds of individuals. It says the group's interests are very broad: politically active people, people in prominent posts, researchers, journalists, public administration and uniformed services staff, and people linked to them by family or friendship. Known harm in 2024: athlete data stolen from a Polish national sports body and published. Known harm to individuals: stolen mailbox logins that give access to messages, contact lists and linked accounts. CERT Polska does not give counts of taken mailboxes.
CERT Polska, 5 June 2025: advised all Roundcube users to update to the current version, to check network logs and received mail for the listed indicators, and, for targeted organizations, to reset passwords, check user activity and remove installed Service Workers. It asked all Polish entities that got similar messages to report to the CSIRT for their group: CSIRT NASK, CSIRT MON or CSIRT GOV.
CERT Polska, 12 June 2026: published the Gmail analysis, with domains and example messages, and urged readers to learn how to use e-mail and social media safely and to report suspicious sites through its incident form.
CSIRT NASK, March 2025: monitored the election related campaigns together with the NASK Information Protection Division of Cyberspace.
The group's aim is the mailbox itself: messages, contacts and linked accounts. Treat private and webmail accounts of staff in sensitive roles as part of your risk.
The 2026 pages ask for the second code and use it at once. App codes and text codes are not enough against this kind of page; hardware keys or passkeys are stronger.
The June 2025 exploit worked on installations that were not updated, and opening the mail was enough. Webmail needs the same patch discipline as other internet services.
The group reused real messages from taken accounts, so one stolen mailbox helps attack the next organization.
CERT Polska relies on reports from people and organizations to track new domains almost daily. Reporting quickly helps others.
Stolen data was published with doping claims; election material was promoted through ads. Plan communication for the day stolen data appears.
If you run Roundcube or any webmail, confirm it is on a current release and check whether CVE-2024-42009 was fixed.
In webmail sessions, check whether unknown Service Workers are registered in staff browsers after a suspicious mail.
Check which staff and board members still use text or app codes, and move high risk people to hardware keys or passkeys.
Offer guidance to politically exposed or senior staff on securing their personal Gmail and other mail accounts.
Watch for new domains that copy your mail provider or brand, especially with unusual endings.
Check what a stolen mailbox would reveal, and limit shared contact lists and old attachments.
Make sure staff know where to report a suspicious mail and that you tell the national CSIRT when you see a campaign.
CERT Polska tracks it as a cluster it attributes to APT activity. It says published work links the group to the Belarusian government and other sources point to Russian intelligence. It is also called Ghostwriter.
CERT Polska said in July 2022 that Ghostwriter first named the operation, and that the two names are now used interchangeably.
Opening the e-mail exploited CVE-2024-42009 and installed a Service Worker that copied genuine logins to an attacker server, according to CERT Polska.
CERT Polska says that since March 2026 the group has run phishing campaigns against Gmail users, and that it also phishes the second login factor.
Politically active people, people in prominent posts, researchers, journalists, public administration and uniformed services staff, and people close to them, according to CERT Polska.
CERT Polska says stolen data about Polish athletes was published on a Telegram channel linked to the group after the Paris Olympics, together with doping disinformation.
Take control of your digital security with an exclusive demo of our AI-driven threat intelligence platform.