Skip to content

QEMU

Bug bounty on Sorona qemu.org

Accepting reports open-source
Top reward
$1,500
Top severity paid
Critical
Assets in scope
11
Public CVEs
60
Last change
No changes yet

At a glance

What to know before you start testing.

Pays from $1,500
for critical bugs.
Partial safe harbor
Some legal protection; read the terms.
Publish after a fix
Coordinate the timing with the company.

Rewards

What Sorona lists as the payout for each severity.

Severity Reward
Critical $1,500+
High $700+
Medium $300+
Low $100 $100 / 1 month Claude Max
Safe harbor
Partial Sorona will not pursue legal action against you for good-faith research submitted to us in line with this policy. We cannot speak for the upstream project, its vendor, or any third party, and we do not grant authorization on their behalf.

Summarised with AI from the program policy on Sorona, read 12 Nov 2026, and checked word for word against it. Read the original

Scope

11 assets you may test, 6 excluded.

Web 1

  • https://www.qemu.org

Products and areas 10

  • Device passthrough and the IOMMU-facing paths, including vhost and vhost-user backends
  • Live migration, including the state one host deserialises from another
  • QEMU's security process splits by use case rather than by component. The virtualization use case is what it triages: a guest that should not be able to reach the host, and a host that should not be exposed by running one. Everything below sits inside that framing
  • TCG, the translation engine, where guest instruction sequences become host code
  • The block layer and every disk image format QEMU parses, since an image file is frequently supplied by somebody other than the operator
  • The firmware and boot interfaces QEMU presents to a guest, which a guest can drive from its first instruction
  • The host-facing services QEMU exposes: NBD, VNC, SPICE and the character device backends
  • The QMP and HMP monitor interfaces and anything a management layer drives through them
  • The whole system-mode device model set: network, block, SCSI, USB, display, audio, character and input devices, emulated and paravirtualised alike
  • virtio in full, including the shared-memory ring handling that every modern guest drives
Out of scope (6)
  • Findings that depend on a long embargo, since the process refuses arbitrary embargo requests
  • Guest-visible crashes and misbehaviour that stay inside the guest without reaching the host or another tenant
  • Operator configuration decisions, including deliberately exposing a monitor socket or a VNC server with no authentication
  • qemu-user emulation, which is the non-virtualization case in practice and therefore the weakest part of this scope
  • The host kernel, KVM itself and the management stack around QEMU, all of which belong to their own projects
  • The non-virtualization use case, which the process states is generally ineligible. Reports there are converted into public bugs

Rules

What the policy asks of you, and the bug types it won't reward, quoted from Sorona's policy.

While testing

  • Submit through sorona.xyz.
  • File a GitLab work item on the QEMU project and mark it confidential before you submit.
  • Email disclosure is no longer accepted. There is no security address to write to
  • Frame the finding as the virtualization use case.
  • Expect a short embargo.
  • Low-severity issues may be published with no patch available, so do not plan a coordinated release around one

Not accepted

  • The non-virtualization use case, which the process states is generally ineligible
  • qemu-user emulation, which is the non-virtualization case in practice
  • Guest-visible crashes and misbehaviour that stay inside the guest without reaching the host or another tenant
  • Operator configuration decisions, including deliberately exposing a monitor socket or a VNC server with no authentication
  • The host kernel, KVM itself and the management stack around QEMU
  • Findings that depend on a long embargo

Summarised with AI from the program policy on Sorona, read 12 Nov 2026, and checked word for word against it. Read the original

Changes

Tracked since 11 Nov 2026. Last checked 1 hour ago.

No changes yet. When the rewards or scope move, you will see it here and in the RSS feed.

Public CVEs

Published vulnerabilities in this vendor's products, from the NVD.

Search NVD
CVESeverityPublished
CVE-2025-54567 Medium 4.2 25 Jul 2025
CVE-2025-54566 Medium 4.2 25 Jul 2025
CVE-2024-7730 High 7.4 14 Nov 2024
CVE-2024-3447 Medium 6.0 14 Nov 2024
CVE-2024-6519 High 8.2 21 Oct 2024
CVE-2024-8354 Medium 5.5 19 Sep 2024
CVE-2024-6505 Medium 6.8 5 Jul 2024
CVE-2024-3567 Medium 5.5 10 Apr 2024
CVE-2024-24474 High 8.8 20 Feb 2024
CVE-2024-26328 Medium 6.0 19 Feb 2024
CVE-2024-26327 Medium 5.3 19 Feb 2024
CVE-2023-6683 Medium 6.5 12 Jan 2024
CVE-2023-6693 Medium 4.9 2 Jan 2024
CVE-2023-2861 Medium 6.0 6 Dec 2023
CVE-2023-5088 Medium 6.4 3 Nov 2023
CVE-2023-2680 High 7.5 13 Sep 2023
CVE-2023-3255 Medium 6.5 13 Sep 2023
CVE-2023-3301 Medium 5.6 13 Sep 2023
CVE-2023-42467 Medium 5.5 11 Sep 2023
CVE-2020-24165 High 8.8 28 Aug 2023
CVE-2022-36648 Critical 10.0 22 Aug 2023
CVE-2023-40360 Medium 5.5 14 Aug 2023
CVE-2023-4135 Medium 6.0 4 Aug 2023
CVE-2023-3180 Medium 6.0 3 Aug 2023
CVE-2023-3019 Medium 6.0 24 Jul 2023
CVE-2023-1386 Low 3.3 24 Jul 2023
CVE-2023-3354 High 7.5 11 Jul 2023
CVE-2023-0664 High 7.8 29 Mar 2023
CVE-2023-1544 Medium 6.0 23 Mar 2023
CVE-2023-0330 Medium 5.3 6 Mar 2023
CVE-2022-4144 Medium 6.5 29 Nov 2022
CVE-2022-4172 Medium 6.5 29 Nov 2022
CVE-2022-3872 High 8.6 7 Nov 2022
CVE-2022-3165 Medium 6.5 17 Oct 2022
CVE-2014-0144 High 8.6 29 Sep 2022
CVE-2014-0147 Medium 6.2 29 Sep 2022
CVE-2014-0148 Medium 5.5 29 Sep 2022
CVE-2022-2962 High 7.8 13 Sep 2022
CVE-2022-0358 High 7.8 29 Aug 2022
CVE-2022-0216 Medium 4.4 26 Aug 2022
CVE-2021-3735 Medium 4.4 26 Aug 2022
CVE-2021-3929 High 8.2 25 Aug 2022
CVE-2021-4158 Medium 6.0 24 Aug 2022
CVE-2020-14394 Low 3.2 17 Aug 2022
CVE-2022-35414 High 8.8 11 Jul 2022
CVE-2021-3611 Medium 6.5 11 May 2022
CVE-2021-3750 High 8.2 2 May 2022
CVE-2021-4206 High 8.2 29 Apr 2022
CVE-2021-4207 High 8.2 29 Apr 2022
CVE-2021-20295 Medium 6.5 1 Apr 2022
CVE-2022-1050 High 8.8 29 Mar 2022
CVE-2021-3582 Medium 6.5 25 Mar 2022
CVE-2021-3748 High 7.5 23 Mar 2022
CVE-2022-26353 High 7.5 16 Mar 2022
CVE-2022-26354 Low 3.2 16 Mar 2022
CVE-2021-20257 Medium 6.5 16 Mar 2022
CVE-2021-3638 Medium 6.5 3 Mar 2022
CVE-2021-3607 Medium 6.0 24 Feb 2022
CVE-2021-3608 Medium 6.0 24 Feb 2022
CVE-2021-3947 Medium 5.5 18 Feb 2022

Original policy

Copied from sorona.xyz on 12 Nov 2026. The summaries on this page are made from it; the live page may have changed since.

Skip to content What this pays This project runs no bounty of its own. sorona.xyz pays you directly on qualified, reproduced findings. Severity │ sorona.xyz pays Critical │ $1,500+ High │ $700+ Medium │ $300+ Low │ $100 / 1 month Claude Max Submit a vulnerability CVE history 343 Last 12 months Memory safety 45 % Denial of service 25 % Privilege escalation 20 % Info disclosure 10 % Latest advisories Loading the feed Overview QEMU is the emulation and virtualization layer under nearly all of Linux virtualization. It sits beneath KVM in libvirt, Proxmox, oVirt and OpenStack, runs the Android emulator, and is what makes multi-architecture container builds work in CI. Its device models parse data a guest controls, so a bug in the network, USB, SCSI or display emulation is a route from inside a virtual machine onto the host running it. QEMU carries by far the heaviest CVE traffic of anything in this toolchain category. The project runs no bounty and takes reports as confidential GitLab issues, so sorona.xyz pays you directly on qualified, reproduced findings. In scope QEMU's security process splits by use case rather than by component. The virtualization use case is what it triages: a guest that should not be able to reach the host, and a host that should not be exposed by running one. Everything below sits inside that framing The whole system-mode device model set: network, block, SCSI, USB, display, audio, character and input devices, emulated and paravirtualised alike virtio in full, including the shared-memory ring handling that every modern guest drives The block layer and every disk image format QEMU parses, since an image file is frequently supplied by somebody other than the operator Live migration, including the state one host deserialises from another The QMP and HMP monitor interfaces and anything a management layer drives through them TCG, the translation engine, where guest instruction sequences become host code The host-facing services QEMU exposes: NBD, VNC, SPICE and the character device backends Device passthrough and the IOMMU-facing paths, including vhost and vhost-user backends The firmware and boot interfaces QEMU presents to a guest, which a guest can drive from its first instruction Out of scope The non-virtualization use case, which the process states is generally ineligible. Reports there are converted into public bugs qemu-user emulation, which is the non-virtualization case in practice and therefore the weakest part of this scope Guest-visible crashes and misbehaviour that stay inside the guest without reaching the host or another tenant Operator configuration decisions, including deliberately exposing a monitor socket or a VNC server with no authentication The host kernel, KVM itself and the management stack around QEMU, all of which belong to their own projects Findings that depend on a long embargo, since the process refuses arbitrary embargo requests Most-valued findings Guest-to-host escape out of a device model. This is the class the security process exists for and where QEMU's CVE volume concentrates, in network, USB, SCSI and display emulation Memory corruption in virtio or in the block layer reachable from data a guest writes into a shared ring or an image file Host memory disclosure into a guest, which on shared infrastructure crosses a tenant boundary rather than only a process boundary Migration state deserialisation defects, where one host acts on a structure another host produced Anything reachable through NBD, VNC or SPICE from the network rather than only from a guest Denial of service that takes down the host or the other guests on it, as opposed to only the guest that caused it Program rules Submit through sorona.xyz. This project runs no bounty of its own; we validate, coordinate disclosure with the maintainers, and pay you directly. File a GitLab work item on the QEMU project and mark it confidential before you submit. The process is explicit about that order Email disclosure is no longer accepted. There is no security address to write to Frame the finding as the virtualization use case. Reports outside it are generally ineligible and get reopened as ordinary public bugs qemu-user is the weak part of this scope. A user-mode emulation finding is likely to be handled as an ordinary bug however real it is Expect a short embargo. The project holds confidentiality only as long as it takes to build a patch and allocate a CVE, and generally refuses arbitrary embargo requests Low-severity issues may be published with no patch available, so do not plan a coordinated release around one No response time is published and the project pays nothing for a report Disclosure Coordinated through a confidential GitLab work item. QEMU runs a deliberately minimal embargo, holding confidentiality only as long as it takes to develop a patch and allocate a CVE, and it declines arbitrary embargo requests. Low-severity issues can be disclosed publicly with no patch available. No response time is published, so treat the timeline as the project's call rather than yours. Safe harbor Sorona will not pursue legal action against you for good-faith research submitted to us in line with this policy. We cannot speak for the upstream project, its vendor, or any third party, and we do not grant authorization on their behalf. You are responsible for understanding and following the laws that apply to you, wherever you are, and for staying within any scope the project has defined. Submit a vulnerability