Back Gbhackers

Trojanized PyPI AI Proxy Steals Claude Prompt, Exfiltrates Data

Gbhackers •Mayura Kathir • April 6, 2026

A malicious PyPI package, hermes-px, that masquerades as a “Secure AI Inference Proxy” while secretly stealing user prompts and abusing a private university AI service. Marketed as an OpenAI-compatible, Tor-routed proxy requiring no API keys, the package actually hijacks a Tunisian university’s internal AI endpoint, injects a stolen Anthropic Claude system prompt, and exfiltrates every […]

The full story

This article is one source in a clustered incident — the cluster page carries the summary, timeline and every other outlet covering it.

Malicious PyPI Package Steals Data via Trojanized AI Proxy A malicious Python package named hermes-px has been identified on the Python Package Index (PyPI), disguised as a 'Secure AI Inference Proxy.' This package is designed to route AI requests through the Tor network, falsely claiming to protect user anonymity. However, it hijacks a private university's internal AI endpoint in Tunisia and exfiltrates sensitive user prompts, including a stolen prompt from Anthropic's Claude AI system. The attack method involves the covert collection of data while users believe they are using a secure service. The full scope of the impact is still being assessed, but the incident raises significant concerns about the security of third-party packages in the Python ecosystem. As of now, there are no reported patches or mitigation strategies available for this specific threat. 3 sources 62

Extracted Entities

Trojanized PyPI AI Proxy Steals Claude Prompt, Exfiltrates Data

The full story

Extracted Entities

Attack Types (2)

MITRE ATT&CK (1)

Platforms (3)