The Hades campaign is a supply chain attack involving 37 malicious wheel artifacts across 19 PyPI packages. The compromised packages deploy a *-setup.pth file that executes automatically on Python interpreter startup, downloading the Bun JavaScript runtime and running an obfuscated JavaScript stealer. This stealer harvests a broad set of developer and CI/CD credentials, including tokens and configuration files for GitHub, npm, PyPI, cloud providers, and container platforms. The campaign is a variant of the Shai-Hulud / Miasma malware lineage, distinguished by new persistence mechanisms and AI evasion techniques such as prompt injection to mislead LLM-based package analysis tools. The malware avoids execution on systems with a Russian locale and attempts lateral movement and propagation via compromised developer environments. The attack also targets computational biology-related packages with a different infection vector embedded in __init__.py files. The campaign demonstrates the risks of trusted package repositories being abused to execute malicious code automatically upon package installation without explicit user invocation.

The attack results in the theft of a wide range of sensitive developer credentials and secrets, including tokens and keys for GitHub, npm, PyPI, cloud platforms (AWS, GCP, Azure), Kubernetes, Docker, Vault, and CI/CD systems. This can lead to unauthorized access to developer environments, source code repositories, cloud infrastructure, and continuous integration pipelines. The malware also installs a destructive wiper service that deletes user data if stolen GitHub tokens are revoked. The campaign enables lateral movement within developer networks and propagation of trojanized packages, increasing the risk of widespread compromise in affected ecosystems.

Patch status is not yet confirmed — check the vendor advisory for current remediation guidance. No official fix or patch information is provided in the source content. Users should avoid installing or updating the identified malicious PyPI packages until further notice. Review and revoke any exposed credentials or tokens that may have been compromised. Monitor for unusual activity related to GitHub tokens and developer environments. Employ additional verification and security controls around package dependencies and supply chain integrity. Given the malware's evasion techniques, rely on multiple detection methods beyond AI-based package analysis tools.