PyPI Blocks Inbox.ru Domains After 1,500+ Fake Package Uploads

The Python Package Index (PyPI) has implemented an administrative block on the inbox.ru email domain, prohibiting its use for new user registrations and as additional verification addresses.

This action stems from a recent campaign that exploited the domain to create over 250 fraudulent accounts, which in turn uploaded more than 1,500 empty projects.

These bogus packages caused widespread confusion among end-users, strained PyPI’s resources, and raised alarms about potential security vulnerabilities.

PyPI administrators swiftly responded by purging all implicated projects from the repository and disabling the associated accounts, underscoring the platform’s commitment to maintaining integrity in the open-source Python ecosystem.

Administrative Crackdown on Spam Campaign

The incident highlights the delegated security model inherent in PyPI’s registration process, where users can employ any valid email address, thereby relying on the email provider’s authentication mechanisms.

To mitigate risks from transient or malicious actors, PyPI leverages the disposable-email-domains list to bar registrations from known disposable services, supplemented by an internal blocklist curated by administrators based on observed abuse patterns.

This is not an isolated measure; a prior instance involved restricting a prominent email provider following similar exploits, as detailed in previous announcements.

The inbox.ru blockade serves as a targeted escalation, aimed at preventing the domain’s misuse in automated account creation scripts that could facilitate typosquatting or other deceptive practices.

Technical Details of the Attack

The campaign unfolded systematically with the creation and verification of an initial user account, complete with two-factor authentication (2FA) setup and API token provisioning indicating a premeditated effort to establish persistent access.

This was followed by bursts of activity: 46 accounts on June 11 over three hours, and a massive 207 on June 24 within four hours.

Project uploads commenced on June 26, escalating rapidly with peaks of 295 on June 27 and 740 on June 30, culminating in a total of 1,525 packages by July 11.

These projects were devoid of functional code, suggesting they were placeholders designed to exploit PyPI’s namespace, potentially hijacking popular entrypoints like command-line interfaces that do not necessitate matching the project’s canonical name.

This pattern aligns with a reconnaissance or “dry run” for larger-scale attacks, where attackers confirm scalability before deploying malware or engaging in slopsquatting a variant of typosquatting where slight misspellings deceive users.

PyPI was first alerted on July 8, 2025, via a user report involving a large language model (Claude 3.5 Sonnet) erroneously recommending installation of a non-existent package, illustrating the perils of unverified third-party advice from AI systems or online sources.

Users are urged to meticulously validate package names prior to installation, avoiding blind copy-pasting that could lead to dependency confusion or supply chain compromises.

In broader terms, this event reinforces the evolving threats to package repositories, where adversaries leverage automated tooling for mass account generation and namespace pollution.

PyPI’s proactive stance, including real-time monitoring and community-driven reporting, exemplifies best practices in repository governance, but it also calls for enhanced user vigilance and possibly advanced detection mechanisms like machine learning-based anomaly identification to preempt future incursions.

As Python’s central hub for over 500,000 projects, maintaining PyPI’s trustworthiness is paramount to the global developer community.

Get Free Ultimate SOC Requirements Checklist Before you build, buy, or switch your SOC for 2025 - Download Now