Executive Summary
In September 2025, the Python Software Foundation discovered that the GhostAction supply chain attack led to the exfiltration of thousands of sensitive access tokens—including PyPI, npm, DockerHub, and AWS keys—from compromised GitHub repositories via malicious GitHub Actions workflows. Although attackers obtained over 3,300 secrets by modifying open-source project workflows to funnel credentials to remote servers, swift response from security teams and open-source maintainers prevented the abuse of stolen tokens on PyPI. The PyPI team proactively invalidated all impacted tokens and urged affected maintainers to adopt short-lived Trusted Publishers tokens and review account security.
This incident highlights the increasing sophistication and scope of supply chain attacks targeting developer ecosystems and CI/CD pipelines, underscoring the urgent need for automated detection, better secrets management, and cross-ecosystem collaboration. The prevalence of such attacks demonstrates an evolving threat landscape exploiting automation and cloud-based tooling.
Why This Matters Now
This incident emphasizes urgent gaps in CI/CD pipeline security and growing attacker focus on developer supply chains. With attackers leveraging automated workflows to steal credentials at scale, organizations face heightened risks to both proprietary and open-source software, making real-time detection and secrets management more critical than ever.
Attack Path Analysis
Attackers compromised GitHub repositories by injecting malicious workflows that harvested sensitive secrets, including PyPI publishing tokens. Using the stolen tokens, they attempted access escalation and targeted multiple package ecosystems. The attackers could pivot across cloud and SaaS environments using these credentials. Remote servers were controlled to receive exfiltrated tokens from workflows, enabling covert communications. Exfiltration occurred as secrets were sent to attacker-controlled infrastructure. Ultimately, impact was contained due to proactive token invalidation, but the potential included supply chain compromise and ecosystem disruption.
Kill Chain Progression
Initial Compromise
Description
Adversaries exploited weaknesses in GitHub Actions workflows to insert malicious jobs, enabling access to secrets stored within repository configurations.
Related CVEs
CVE-2025-30154
CVSS 9.8Compromised GitHub Action 'reviewdog/action-setup@v1' allowed exfiltration of secrets via workflow logs.
Affected Products:
reviewdog action-setup – v1
Exploit Status:
exploited in the wildCVE-2025-54416
CVSS 9.1tj-actions/branch-names GitHub Action vulnerable to command injection via crafted branch names or tags.
Affected Products:
tj-actions branch-names – < 2.0.0
Exploit Status:
proof of conceptReferences:
CVE-2024-42471
CVSS 8.8GitHub's actions/artifact vulnerable to arbitrary file write via path traversal in artifact extraction.
Affected Products:
GitHub actions/artifact – < 2.1.2
Exploit Status:
no public exploitCVE-2020-15228
CVSS 8GitHub Actions' set-env and add-path commands vulnerable to environment variable injection.
Affected Products:
GitHub Actions Runner – < 2.273.1
Exploit Status:
exploited in the wild
MITRE ATT&CK® Techniques
Adversary-in-the-Middle: Web Protocols
Unsecured Credentials: Credentials in Files
Modify Authentication Process: Pluggable Authentication Modules
Exfiltration Over Web Service: Exfiltration to Cloud Storage
Command and Scripting Interpreter: Python
Valid Accounts
Container Administration Command
Compromise Software Supply Chain: Compromise Software Dependencies and Development Tools
Potential Compliance Exposure
Mapping incident impact across multiple compliance frameworks.
PCI DSS v4.0 – Segregation of Duties for Development Environments
Control ID: 12.3.3
NYDFS 23 NYCRR 500 – Cybersecurity Policy
Control ID: 500.03
DORA (Digital Operational Resilience Act) – ICT Risk Management Framework
Control ID: Art. 9
CISA Zero Trust Maturity Model (ZTMM) 2.0 – Credential and Secrets Governance
Control ID: Identity Pillar: Credentials and Secrets Management
NIS2 Directive – Cybersecurity Risk Management and Reporting
Control ID: Article 21
Sector Implications
Industry-specific impact of the vulnerabilities, including operational, regulatory, and cloud security risks.
Computer Software/Engineering
PyPI supply chain attack directly targets software development workflows, compromising package publishing tokens and GitHub Actions, threatening code integrity and deployment security.
Information Technology/IT
GhostAction campaign exploits CI/CD pipelines and package management systems, requiring immediate token rotation and adoption of short-lived Trusted Publishers for security.
Financial Services
Stolen AWS access keys and database credentials in supply chain attack pose compliance risks under PCI DSS, threatening customer data protection.
Computer/Network Security
Multi-ecosystem token compromise across PyPI, npm, DockerHub demonstrates sophisticated supply chain vulnerabilities requiring enhanced egress security and anomaly detection capabilities.
Sources
- PyPI invalidates tokens stolen in GhostAction supply chain attackhttps://www.bleepingcomputer.com/news/security/pypi-invalidates-tokens-stolen-in-ghostaction-supply-chain-attack/Verified
- Token Exfiltration Campaign via GitHub Actions Workflowshttps://blog.pypi.org/posts/2025-09-16-github-actions-token-exfiltration/Verified
- PyPI Invalidates Tokens Stolen in GhostAction Supply Chain Attackhttps://cirt.gy/article/al2025_39-pypi-invalidates-tokens-stolen-in-ghostaction-supply-chain-attack-september-23rd-2025/Verified
- GitHub supply chain attack sees thousands of tokens and secrets stolen in GhostAction campaignhttps://www.techradar.com/pro/security/github-supply-chain-attack-sees-thousands-of-tokens-and-secrets-stolen-in-ghostaction-campaignVerified
Frequently Asked Questions
Cloud Native Security Fabric Mitigations and ControlsCNSF
Zero Trust network segmentation, granular egress policy enforcement, east-west traffic controls, and centralized visibility—aligned with CNSF capabilities—would have limited attackers' ability to move, exfiltrate secrets, or escalate access. Proactive detection and inline policy enforcement reduce the blast radius of automated supply chain attacks.
Control: Zero Trust Segmentation
Mitigation: Reduces attacker access to critical secrets by isolating workloads and enforcing least privilege between automation systems.
Control: Multicloud Visibility & Control
Mitigation: Detects suspicious use of privileged tokens and highlights anomalous authentication behavior.
Control: East-West Traffic Security
Mitigation: Prevents unauthorized connections between services and environments, stopping lateral movement.
Control: Cloud Firewall (ACF)
Mitigation: Blocks malicious external communications and DNS-based C2 by filtering unauthorized destinations.
Control: Egress Security & Policy Enforcement
Mitigation: Prevents unauthorized data egress and detects secret exfiltration attempts.
Early detection and response limits opportunity for malicious artifacts to propagate.
Impact at a Glance
Affected Business Functions
- Software Development
- Continuous Integration/Continuous Deployment (CI/CD)
Estimated downtime: 7 days
Estimated loss: $500,000
Exfiltration of over 3,300 secrets, including API tokens and credentials, potentially compromising multiple software repositories and associated services.
Recommended Actions
Key Takeaways & Next Steps
- • Enforce zero trust segmentation to ensure CI/CD workflows only access minimal secrets and approved services.
- • Implement granular outbound (egress) policies to restrict where automation and workloads can connect, blocking data exfiltration paths.
- • Centralize multicloud and repo visibility to detect anomalous workflow behavior and unauthorized use of publishing tokens.
- • Apply east-west workload traffic controls to stop lateral movement between services, clouds, and CI/CD environments.
- • Enable continuous threat detection and automated response for real-time alerting on suspicious workflow or credential activity.



