Major

Fortune

Million

<h2 id="key-takeaways">Key Takeaways</h2>
<ul>
 <li>Internal source code from Claude Code, Anthropic’s AI coding assistant, was unintentionally made public</li>
 <li>Approximately 1,900 files containing 512,000 lines of code were exposed</li>
 <li>A social media post linking to the leaked code surpassed 30 million views</li>
 <li>The company confirms no user data or authentication credentials were compromised</li>
 <li>Anthropic has experienced two separate data exposure incidents within the same week</li>
</ul>
<p>On Tuesday, Anthropic publicly acknowledged an unintended disclosure of proprietary source code belonging to Claude Code, its artificial intelligence-driven development tool. Company representatives characterized the incident as stemming from a “release packaging issue caused by human error, not a security breach.”</p>
<div class="embed-twitter">
 <blockquote class="twitter-tweet">
  <p>Claude code source code has been leaked via a map file in their npm registry!</p>
  <p>Code: <a href="https://t.co/jBiMoOzt8G" rel="noopener noreferrer nofollow">https://t.co/jBiMoOzt8G</a><a href="https://t.co/rYo5hbvEj8" rel="noopener noreferrer nofollow">pic.twitter.com/rYo5hbvEj8</a></p>
  <p>— Chaofan Shou (@Fried_rice) <a href="https://twitter.com/Fried_rice/status/2038894956459290963?ref_src=twsrc%5Etfw" rel="noopener noreferrer nofollow">March 31, 2026</a></p>
 </blockquote>
</div>
<p>Cybersecurity professionals analyzing the exposure reported that nearly 1,900 individual files totaling approximately 512,000 lines of code became accessible. Given that Claude Code operates within developer environments where it interacts with confidential information, security specialists expressed significant apprehension about the implications.</p>
<p>A social media message on X containing a direct link to the exposed code rapidly gained traction online. Within hours of appearing during the early Tuesday morning, the post had already accumulated more than 30 million impressions.</p>
<p>Software engineers immediately began analyzing the released code to gain insights into Claude Code’s operational mechanics and Anthropic’s future development roadmap. Meanwhile, security professionals highlighted potential exploitation vectors that malicious actors might leverage with access to this information.</p>
<p>AI-focused cybersecurity company Straiker published analysis suggesting that threat actors could now examine Claude Code’s internal data processing architecture. Their assessment indicated this knowledge could enable adversaries to engineer persistent exploits that maintain presence throughout extended sessions, essentially creating backdoor access.</p>
<h2 id="back-to-back-security-events">Back-to-Back Security Events</h2>
<p>This incident represents the second problematic disclosure for <a href="https://coincentral.com/xstocks-and-fundrise-tokenize-private-tech-fund-bringing-spacex-and-anthropic-on-chain/" rel="noopener noreferrer nofollow">Anthropic</a> in under seven days. Fortune previously disclosed that the organization had mistakenly granted public access to thousands of internal documents.</p>
<p>Among those documents was an unpublished blog entry detailing a forthcoming AI system referenced internally under the codenames “Mythos” and “Capybara.” The draft reportedly acknowledged potential cybersecurity vulnerabilities associated with the model.</p>
<p>Anthropric announced plans to implement additional safeguards against similar future occurrences. The organization emphasized that neither incident involved exposure of customer information or authentication credentials.</p>
<h2 id="claude-code-s-market-position">Claude Code’s Market Position</h2>
<p>Anthropric made Claude Code publicly available in May of the previous year. The platform assists programmers with feature development, debugging, and workflow automation.</p>
<p>Adoption has accelerated substantially. By February, the product had achieved an annualized revenue run rate exceeding $2.5 billion.</p>
<p>This commercial success has intensified competitive pressure across the sector. OpenAI, Google, and xAI have all committed significant resources toward developing comparable coding assistance platforms to challenge <a href="https://coincentral.com/anthropics-secret-weapon-just-got-leaked-meet-claude-mythos/" rel="noopener noreferrer nofollow">Claude</a> Code’s market position.</p>
<p>Founded in 2021 by former OpenAI leadership and research personnel, Anthropic has built its reputation primarily around its Claude AI model series.</p>
<p>A company representative confirmed that Anthropic is implementing procedural changes designed to prevent recurrence of such disclosure events.</p>
<p>The post <a href="https://blockonomi.com/anthropic-exposes-half-a-million-lines-of-claude-code-in-major-leak/" rel="noopener noreferrer nofollow">Anthropic Exposes Half a Million Lines of Claude Code in Major Leak</a> appeared first on <a href="https://blockonomi.com" rel="noopener noreferrer nofollow">Blockonomi</a>.</p>

Anthropic Exposes Half a Million Lines of Claude Code in Major Leak

Table of Contents

Blockonomi

Other articles published on Apr 1, 2026

Join our free newsletter for daily crypto updates!