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
CVE-2026-97229-GitLab-CI-CD-Regex-Compiler-RCE — Technical investigation of CVE-2026-97229 (GitLab CI/CD regex RCE): local patched GitLab lab, CI Lint API checks, version auditing, and defensive guidance. No working exploit. | Kitploit
Tools/GitHubGitHub/ampedwastaken/cve-2026-97229-gitlab-ci-cd-regex-compiler-rce
Defensive ToolsVulnerability AnalysisConfiguration AuditingInformation GatheringDevSecOpsPapers & ResearchLearning & Education
GitHub
Share
ampedwastaken/cve-2026-97229-gitlab-ci-cd-regex-compiler-rce

CVE-2026-97229-GitLab-CI-CD-Regex-Compiler-RCE

Technical investigation of CVE-2026-97229 (GitLab CI/CD regex RCE): local patched GitLab lab, CI Lint API checks, version auditing, and defensive guidance. No working exploit.

View Repository
1161 month agoNot yet reviewed

CVE-2026-97229 — GitLab CI/CD Regex Compiler RCE

Technical investigation | Real GitLab CE | CVSS 10.0 | 2026-10-13

Exploit disclosure status: GitLab has confirmed the vulnerability, but the precise trigger, affected implementation, and RCE chain are not publicly documented as of 13 October 2026. This repository is not a working RCE PoC. It uses a real, patched GitLab server and documented CI Lint API to study the legitimate CI/CD regex interface. It does not pretend normal lint requests exploit the vulnerability.

The vulnerability

On 12 October 2026, GitLab published fixes for CVE-2026-97229, an integer overflow / insufficient regex validation issue impacting GitLab CE/EE. GitLab says a specially crafted regular expression in a CI/CD configuration could, under certain conditions, allow unauthenticated arbitrary code execution on the GitLab server.

Reported by: joaxcar, credited by GitLab.

GitLab CE/EE releaseAffected versionsGitLab's current fix
19.2.xBefore 19.2.819.2.8
19.3.xBefore 19.3.519.3.5
19.4.xBefore 19.4.219.4.2

Severity: CVSS 3.1 10.0 (AV:N/AC:L/PR:N/UI:N/S:C/C:H/I:H/A:H). Self-managed operators should upgrade immediately. GitLab.com was already patched; GitLab Dedicated customers were told no action was required for this issue.

Source: GitLab's official critical patch release.

What is real in this repository?

ComponentWhat it actually does
compose.yamlStarts official gitlab/gitlab-ce:19.4.2-ce.0 on 127.0.0.1:8929
fixtures/legitimate-ci-regex.ymlReal GitLab .gitlab-ci.yml rules using regex operators
tools/ci_lint.pyPosts that configuration to a local GitLab project-scoped CI Lint API
tools/version_audit.pyOffline version inventory matching the published affected ranges
tools/instance_version.pyReads a running localhost GitLab server’s real version via its Metadata API
docs/RESEARCH.mdPublic sources, known facts, unknowns, and source-review process
docs/DEFENSIVE_GUIDE.mdPatching, validation, and defensive triage
tests/Automated tests for API request construction, parsing and version classification

This replaces an earlier made-up 8-bit arithmetic simulator. No synthetic "vulnerable compiler" is presented as evidence about GitLab.

Reproduce the real (benign) behavior

1. Start a patched GitLab instance

You need a Linux Docker engine with sufficient RAM and disk for GitLab's official omnibus image. Docker Desktop on Windows may have unsupported volume-permission behaviors according to GitLab documentation; a Linux host or Linux VM is preferred.

docker compose up -d
docker compose ps

Open http://127.0.0.1:8929. Initial boot requires provisioning. To retrieve the locally generated root password, consult the container's /etc/gitlab/initial_root_password file. Set up a private local project and create a personal access token with the minimum access necessary for project CI linting.

Never connect this research environment to the public internet. Only port 8929 is published and it is bound to localhost. No external exploit endpoint is used.

2. Test actual GitLab regex validation

GitLab documents regex operators =~ and !~ in CI/CD rules, and the project-scoped CI Lint API at POST /api/v4/projects/:id/ci/lint.

Set GITLAB_TOKEN in your local shell (PowerShell: $env:GITLAB_TOKEN = "YOUR_LOCAL_TOKEN"); do not commit tokens. Supply the numeric ID of your local private project:

python -m tools.ci_lint --project-id 1 --config fixtures/legitimate-ci-regex.yml --output result.json

This uses the real GitLab application and returns its actual JSON lint response (normalized to validity, errors and warnings). It is a legitimate configuration check, not evidence of exploiting this CVE.

3. Classify an installed version

python -m tools.version_audit 19.4.1
python -m tools.version_audit 19.4.2

The first reports affected and exits 2; the second reports fixed_for_this_cve. Unlisted release lines are explicitly outside_documented_ranges, not automatically marked safe.

To inspect your actual local GitLab installation via GitLab’s documented metadata endpoint:

python -m tools.instance_version

This uses your existing GITLAB_TOKEN and requests GET http://127.0.0.1:8929/api/v4/version. This is local inventory, not an exploit test.

4. Run automated tests

python -m unittest discover -s tests -v

Unit tests run without Docker or network access. End-to-end CI lint needs a running local GitLab instance and its token. The optional, manually triggered GitHub Actions workflow only executes offline unit tests, so it cannot claim to verify real GitLab server behavior. Automatic runs are paused while repository Actions jobs are failing.

5. Stop the GitLab instance

docker compose down

Local Docker runtime data are stored in .gitlab-runtime/ and ignored by Git. docker compose down preserves these host directories.

What is not demonstrated

  • No verified vulnerable regex payload or RCE execution primitive
  • No confirmed CVE-specific source file or exact patched hunk
  • No bypass of authentication, and no claim that lint accepts unauthenticated malicious expressions
  • No remote instance scanning, shell execution, or exploitation

GitLab documents that details of security issues are normally made public 90 days after the associated patch release. Research should be updated with attributed code references when verified details emerge.

Evidence and primary sources

  • GitLab 19.4.2 critical patch advisory
  • GitLab CI job rules: regular expressions
  • GitLab CI Lint API
  • GitLab Metadata API
  • GitLab Docker installation
  • CERT-EU advisory

Research integrity: Attribute GitLab and joaxcar. Do not label this repository "working CVE-2026-97229 exploit" until an independently reproducible CVE-specific result exists. It is a practical technical investigation using real vendor software, not an attack tool.

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.

    945
  • 2

    ghostlock-app

    yukonga/ghostlock-app

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

    555
  • 3

    redforge

    cybersecurity91/redforge

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

    350
  • 4

    ARTEX

    baarai/artex

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

    339
  • 5

    looking-glass

    anonghosty/looking-glass

    CSIRT platform that ingests Shadowserver reports into MongoDB and replays them as interactive maps, analysis, findings, advisories and executive dashboards.

    338
  • 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.

    292

Explore all tools

Browse our collection of tools

View all tools →