CodeQL Wall of Fame
Join us in our mission to improve open source security for all
456
vulnerabilities found
with the help of CodeQL
The GitHub Security Lab uses CodeQL to perform
variant analysis, an important technique for identifying new types
of security vulnerabilities of a given class. The Security Lab and
its community shares its knowledge with developers, to benefit
both open source and commercial organizations.
The CodeQL Wall of Fame is a (non-exhaustive) list of
vulnerabilities that the GitHub Security Lab and our community
have found using CodeQL. In most cases these vulnerabilities were
detected as a direct result of a query launch. In other cases,
CodeQL was used to explore the codebase faster and accelerate the
manual audit.
Take action
Want to join us in our mission to improve open source security for all? Choose your own adventure to get started:
Get featured
Did you find a new CVE thanks to CodeQL? Open a submission to
see your work displayed on the CodeQL Wall of Fame.
Secure your code
You are one click away to benefit from the power of CodeQL on
your open source codebase. For private code, contact sales.
Learn CodeQL
Dive into our Capture the Flag challenges designed to sharpen
your abilities while mastering CodeQL.
Featured disclosures
May 22, 2026
Checking out of untrusted code in benchmarks.yml workflow may lead to code execution in privileged runner
April 3, 2026
The PraisonAI repository was vulnerable to code injection in the claude-code-action GitHub Actions workflow (GHSL-2025-093).
December 19, 2025
A code injection vulnerability was identified in the then latest changeset of vets-api's GitHub Actions workflow (ready_for_review.yml), which could have allowed unauthorized code execution during workflow runs.
All advisories
Since March 2020
2026
GHSL-2026-203: Same-second TOCTOU vulnerability in jupyterlab/maintainer-tools update-snapshots-checkout - CVE-2026-84973
The jupyterlab/maintainer-tools update-snapshots-checkout action uses one-second timestamps to reject pull request updates after an authorized comment. An attacker can win a same-second race, causing a privileged workflow to check out and execute attacker-controlled code.
GHSL-2026-199: Code execution in DataDog/trivy API Diff Check workflow
The API Diff Check workflow invokes a pull request-controlled local action in a privileged pull_request_target job. An external contributor can execute arbitrary code with a write-scoped GITHUB_TOKEN and access the ORG_REPO_TOKEN secret.
GHSL-2024-200_GHSL-2024-201: Zero click RCE in Uptrain - CVE-2025-27771, CVE-2025-27772
The Uptrain dashboard lacks significant authentication, has an open CORS policy, and is vulnerable to a remote code execution vulnerability. Combining these primitives, an attacker can get zero click remote code execution in the context of the Uptrain host by directing an Uptrain user to a specially crafted website.
GHSL-2024-198_GHSL-2024-199: Zero click RCE in Uptrain - CVE-2025-27621, CVE-2025-27770
The Uptrain dashboard lacks significant authentication, has an open CORS policy, and is vulnerable to a remote code execution vulnerability. Combining these primitives, an attacker can get zero click remote code execution in the context of the Uptrain host by directing an Uptrain user to a specially crafted website.
GHSL-2024-196: Remote Code Execution (RCE) in Fooocus webui - CVE-2025-31114
The Fooocus webui is vulnerable to remote code execution due to unsafe use of eval when processing metadata json. An attacker with access to Foocus webui page may be able to execute arbitrary code on the instance.
GHSL-2025-009: Arbitrary Code Execution in Mesa benchmarks.yml workflow - CVE-2026-29075
Checking out of untrusted code in benchmarks.yml workflow may lead to code execution in privileged runner
GHSL-2024-293: Arbitrary File Read on get_log_file REST API in Apache Doris
The get_log_file REST API allows user with Admin privilege to read arbitrary files on the server
GHSL-2025-093: Code Injection in PraisonAI claude-code-action GitHub Actions Workflow
The PraisonAI repository was vulnerable to code injection in the claude-code-action GitHub Actions workflow (GHSL-2025-093).
2025
GHSL-2025-105: Code injection in vets-api
A code injection vulnerability was identified in the then latest changeset of vets-api's GitHub Actions workflow (ready_for_review.yml), which could have allowed unauthorized code execution during workflow runs.
GHSL-2025-102_GHSL-2025-103: Code injection in acl-anthology
The latest changeset of acl-anthology was vulnerable to code injection in two GitHub Actions workflows, link-to-checklist and print-info, which could have allowed attackers to execute arbitrary commands during CI processes.
GHSL-2025-106: Code Injection in esphome/esphome-docs Github Actions Workflow
A code injection vulnerability exists in the latest main branch of esphome/esphome-docs, where the .github/workflows/component-image.yml Github Actions workflow allows attackers to execute arbitrary code with privileged context. This flaw could enable unauthorized access or compromise of the CI environment.
GHSL-2025-104: Arbitrary command execution within the CI environment in Weaviate
The Weaviate repository was vulnerable to code injection in its Github Actions notifier.yml as of the latest changeset at the time of the review, which potentially could have allowed attackers to execute arbitrary commands within the CI environment.
GHSL-2025-099: Arbitrary code execution in cross-platform-actions/action
The cross-platform-actions/action repository was vulnerable to a code injection issue in its release.yml workflow, affecting the latest changeset at the time of review. This vulnerability could potentially have allowed an attacker to inject and execute arbitrary code within the context of the workflow.
GHSL-2024-325_GHSL-2024-326: Poisoned Pipeline execution via Code Injection and Execution of untrusted code in Actual
The update-vrt.yml workflow is vulnerable to Poisoned Pipeline execution via Code Injection and Execution of untrusted code
GHSL-2025-094: code execution in a privileged context in a GitHub workflow of faststream
The GitHub Actions workflow pr_autoupdate.yaml checks out untrusted code, potentially allowing attackers to execute arbitrary code in a privileged context.