Skip to content

Latest commit

 

History

History
162 lines (110 loc) · 9.36 KB

File metadata and controls

162 lines (110 loc) · 9.36 KB

Bibliography of GitHub Actions Security Research

The Monsters in Your Build Cache – GitHub Actions Cache Poisoning

Read Article
2024 (UPDATE 01/23/25), Blog Post By Adnan Khan.

Problem Addressed

This blog post by Adnan Khan explores a critical security issue in GitHub Actions caching and how attackers can manipulate it to gain unauthorized access to CI/CD workflows. The core issue is cache poisoning, allowing malicious actors to replace legitimate cache entries, leading to privilege escalation and exposure of sensitive data.

Key Takeaways

  • Detailed analysis of GitHub Actions caching mechanisms and potential attack vectors.
  • Step-by-step execution of cache poisoning attacks.
  • Highlights the real-world impact on projects like Angular, Mozilla MDN, and Hyperledger Besu.
  • Introduces the Actions Cache Blasting attack technique.

Potential Improvements

The article thoroughly explains the attack but lacks detailed discussion on mitigation strategies. It could be improved by suggesting stricter cache validation, exploring ways GitHub could enhance security, and providing tools for maintainers to detect cache poisoning risks.

Related Security Tools


How to Secure Your GitHub Actions Workflows with CodeQL

Read Article
Jan 9, 2025, GitHub Security Blog By Alvaro Munoz.

Problem Addressed

This article from GitHub's security blog discusses how CodeQL, GitHub’s static analysis tool, has been enhanced to detect vulnerabilities in GitHub Actions workflows. It addresses workflow misconfigurations and untrusted data flow risks through taint tracking analysis.

Key Takeaways

  • Taint tracking for untrusted data sources, improving security scans.
  • Introduces GitHub Actions as a structured language in CodeQL.
  • Supports Bash script analysis, allowing security checks within shell commands.
  • Detects cache poisoning vulnerabilities, mitigating supply chain attacks.
  • Already in use to secure 75+ major open-source projects and detect 90+ vulnerabilities.

Potential Improvements

While CodeQL’s security enhancements are valuable, improvements could include expanding taint tracking for complex workflows involving multiple repositories, providing automated remediation suggestions, and improving detection of logic-based attacks such as TOCTOU vulnerabilities.

Related Security Tools


I'll Think Twice Before Using GitHub Actions Again

Read Article
2025 (Update 30.01.2025), Blog Post By Nemanja Ninković.

Problem Addressed

This blog post critiques GitHub Actions' usability and scalability issues, especially for monorepos and complex CI/CD pipelines. The author argues that GitHub Actions suffers from inefficiencies, lack of local testing, and YAML complexity, making alternative CI/CD solutions more attractive.

Key Takeaways

  • Inefficient monorepo support, causing merge conflicts.
  • YAML complexity leads to bloated workflows with excessive logic.
  • No local testing support, forcing developers to push untested changes.
  • GitHub is slow to address usability concerns, frustrating users.
  • Alternative CI/CD solutions like GitLab CI/CD, Jenkins, TeamCity, and Dagger provide better scalability.

Potential Improvements

GitHub Actions should allow required checks to dynamically adjust based on executed jobs, improve YAML syntax to reduce redundant logic, enable local workflow execution for easier debugging, and improve communication regarding its roadmap and feature development.


Exploring the Characteristics and Challenges in GitHub Actions and Its Ecosystem

Read Article
Jun 2024, BSc Thesis - University of Gothenburg By Saif Sayed and Kardo Marof.

Problem Addressed

This research explores the challenges within the GitHub Actions ecosystem, analyzing 997 discussion threads and 4.1K repositories to identify common issues like security vulnerabilities, dependency problems, and workflow obsolescence.

Key Takeaways

  • Security vulnerabilities are the most prevalent issue in GitHub Actions workflows.
  • Dependency issues and breaking changes create reliability concerns.
  • Marketplace actions are widely used, but locally maintained actions are preferred in high-complexity repositories.
  • Developers struggle with obsolescence, as outdated actions impact build stability.

Potential Improvements

The study suggests further research into improving security mechanisms for Marketplace Actions, better dependency management practices, and automated solutions for detecting and mitigating breaking changes. It also recommends a deeper qualitative analysis of developer preferences regarding Local vs. Marketplace Actions and a comparison with other CI/CD systems like GitLab CI/CD and Travis CI.

Related Dataset


On the Use of GitHub Actions in Software Development Repositories

Read Article
Oct 2022, IEEE International Conference on Software Maintenance and Evolution - University of Mons By Alexandre Decan, Tom Mens, Pooya Rostami Mazrae and Mehdi Golzadeh.

Problem Addressed

This empirical study analyzes 68K GitHub repositories, showing how GitHub Actions is adopted, which workflows are automated, and the security risks involved.

Key Takeaways

  • 43.9% of repositories use GitHub Actions, making it the dominant CI/CD tool.
  • Reusable actions are heavily used, but security concerns exist.
  • Most workflows focus on CI/CD tasks, with continuous integration and testing being top priorities.
  • Security issues arise from unverified third-party actions, exposing projects to potential threats.

Potential Improvements

The research suggests improving dependency versioning best practices to mitigate security risks, enhancing tooling for workflow management to reduce maintenance overhead, and conducting further research on the long-term evolution of the GitHub Actions ecosystem to address emerging challenges. It also calls for better security practices in reusable actions to minimize the risk of supply chain attacks.

Related Dataset


Research Directions in Software Supply Chain Security

Read Article
2023, ACM Transactions on Software Engineering and Methodology (TOSEM) By Williams et al.

Problem Addressed

This research provides a broad overview of software supply chain security risks, categorizing three major attack vectors: dependencies, build infrastructure, and human factors.

Key Takeaways

  • 96% of software contains open-source dependencies, making them a prime attack target.
  • Build infrastructure is a major attack vector, as seen in the SolarWinds and Log4j incidents.
  • Social engineering targets developers, leading to compromised credentials and injected backdoors.
  • Tools like SCA scanners, SBOMs, and in-toto are essential for mitigating risks.

Potential Improvements

The paper suggests enhancing automated vulnerability detection to improve how dependencies are monitored and updated, developing better tools to analyze and secure build environments, and focusing on the human factor by strengthening identity verification and training against social engineering attacks. It also emphasizes the need for future research into securing AI-generated code and LLM-based development workflows, as these introduce new and largely unexplored risks.

Related Security Tools

  • SCA tools, SBOMs, OpenSSF Scorecard, SLSA, in-toto, TUF

GitHub’s Native Attestation Support

Read Article
2024 (UPDATE frequently), GitHub Documentation By INRIA.

Problem Addressed

Ensuring that software artifacts in GitHub Actions are verifiable and haven’t been tampered with.

Key Takeaways

GitHub provides a web UI ("Attestations" tab) and CLI command (gh attestation verify) for verifying artifact provenance. This helps prevent supply chain attacks.

Potential Improvements

It is unclear how many repositories use attestations and whether GitHub provides a public transparency log for attestations.


GitHub Attestation API

Read Article
Nov 2022, GitHub API Documentation

Problem Addressed

Enables developers to programmatically retrieve and verify artifact attestations.

Key Takeaways

The API allows for automated verification of software artifacts, ensuring that only trusted artifacts are used in production.

Potential Improvements

Fine-grained adoption statistics would help measure the real-world impact of this security feature.