Claude Code Tied to Healthcare Cyberattacks

Anthropic’s Claude Code has drawn heightened scrutiny this week following reports that threat actors used the AI coding agent in cyberattacks targeting healthcare organizations worldwide, while simultaneously seeing broader adoption across the open-source development community. The developments highlight both the security risks and transformative potential of autonomous coding agents as they become embedded in software development workflows.

According to a report from CrowdStrike cited by healthcare publication sources, threat actors deployed Claude Code and other agentic AI systems in campaigns targeting at least 15 healthcare organizations in Argentina, as well as more than 20 healthcare and biotech organizations worldwide. The activity included credential theft and medical record exfiltration, marking what security researchers describe as a growing use of AI agents in offensive cyber operations.

Internal Development Process Revealed

The incidents emerged as Anthropic published an inside account of how Claude Code evolved from an internal command-line tool into the company’s flagship coding agent. The article from the Anthropic Newsroom includes perspectives from researchers, engineers, and early users involved in building the product, offering rare transparency into the development process behind one of the industry’s most advanced AI coding tools.

Separately, reporting from Chinese financial publication Sina Finance indicates that roughly 90% of the Claude Code team’s own code is now AI-generated, according to comments attributed to Anthropic engineer Addy Osmani. The development process has shifted toward emphasizing testing, code review, security checks, evaluation, and human oversight rather than manual code production alone.

Leadership Departures and Enterprise Pressure

The Claude Code project is also facing organizational changes. Senior contributors Thariq Shihipar and Anthony Morris announced their departures from Anthropic, according to a report from Sina Finance. The same report discusses reductions in Claude usage at Microsoft and Meta as those companies promote GitHub Copilot, internal coding tools, and proprietary models, potentially affecting the agent’s enterprise adoption trajectory.

High-Profile Open Source Adoption

Despite the security concerns and internal changes, Claude Code continues to see adoption among prominent developers. Linux creator Linus Torvalds reportedly used Claude Code under his supervision to draft eight experimental Linux kernel changes aimed at addressing a performance issue. The changes had not been merged into the mainline kernel as of the report, underscoring that the work remained subject to human technical review and validation.

The workflow configuration shared by Y Combinator CEO Garry Tan has also gained substantial attention. His open-source Claude Code setup, called “gstack,” uses role-based commands for functions including product management, engineering, design, security, quality assurance, and release management, demonstrating how developers are structuring AI agents for complex, multi-stage development tasks.

Changing Hiring Practices

Anthropic hiring manager Felix Rieseberg reportedly said he stopped using traditional coding interviews because Claude performs better at many coding tasks. The change represents a broader shift toward evaluating problem-solving ability, motivation, and engineering judgment rather than manual coding proficiency in technical interviews.

Implications for Search and Content Discovery

The developments around Claude Code are occurring as AI systems reshape how users discover information online. A sponsored article in USA Today discusses how search-engine optimization and generative-engine optimization are changing as users increasingly receive answers directly from search and AI systems, citing a study that claims 68% of Google searches end without a click.

Key Facts

CrowdStrike reported threat actors used Claude Code in campaigns targeting at least 15 healthcare organizations in Argentina and more than 20 healthcare and biotech organizations globally
Approximately 90% of Claude Code team’s code is AI-generated, according to Anthropic engineer Addy Osmani
Senior Claude Code contributors Thariq Shihipar and Anthony Morris have departed Anthropic
Linus Torvalds used Claude Code to draft eight experimental Linux kernel changes under human supervision
Anthropic hiring manager Felix Rieseberg stopped giving traditional coding interviews, citing Claude’s superior performance on many coding tasks
Microsoft and Meta reportedly reducing Claude usage in favor of GitHub Copilot and proprietary models

Sources

Written by

Mike Gingerich

Author

Author: Mike Gingerich, President of web firm Digital Hill, Co-Founder of TabSite .
Digital and Social Media Marketer, Speaker, and Business Consultant. Part geek, part marketer, total digital junkie! Seeking to add value, make the complex simple, and leave a positive impact.

Follow me on twitter: @mike_gingerich.