
Anthropic Addresses Leak of 3,000 Files from AI Coding Tool's Source Code
Anthropic AI Tool Leak Exposes 500,000 Lines of Code
Anthropic, a leading developer of artificial intelligence (AI) tools, has come under scrutiny after a major code leak revealed internal details of its AI tool, Claude Code. The incident reportedly made around 3,000 files public, along with nearly 500,000 lines of source code across 1,900 files.
According to Anthropic, the leak was caused by human error, specifically a misconfiguration of their content management system. The company confirmed that "some internal source code" became public during a "Claude Code release," adding that no sensitive data was involved. This is not the first incident of its kind, as Anthropic accidentally exposed Claude Code's source code in February 2025, raising concerns about its safeguards.
The leaked code includes a draft post about an upcoming model, known internally as "Mythos" or "Capybara." Experts warn that even though the core AI model was not leaked, the exposed code could still reveal important internal workings. A cybersecurity expert notes that developers may now be able to extract useful insights from the codebase, which could potentially help rivals study the system and create competing tools.
Read also: Bitcoin Price Sinks 6% Below $66,500 Amid ETF Outflows and Institutional Selling
Claude Code is a highly advanced AI tool used by large companies, and it depends not only on AI models but also on a system that guides how the AI works. This "harness" controls behavior and connects the AI to tools, and it is this system that has been leaked. Experts warn that the leak could lead to open-source versions based on the exposed code.
| Model | Number of Lines of Code |
|---|---|
| Claude Code | 500,000 |
| Mythos/Capybara | ? (estimated) |
| Opus | ? (estimated) |
| Sonnet | ? (estimated) |
| Haiku | ? (estimated) |
According to researcher Roy Paz, the leak also hints at a new advanced model, possibly more powerful than current versions. Anthropic currently offers models like Opus, Sonnet, and Haiku, but the new system may go beyond them. Paz notes that the leak suggests that Anthropic is working on a more advanced model, likely stronger than its current offerings.
The incident has caught the internet's attention, with users sharing strong reactions. Some have questioned whether the leak was intentional, while others have pointed out that many breaches are caused by simple human errors, such as forgetting to flip a private toggle.
Read also: Bitcoin's Inflation-Hedging Potential Erodes as Price Falls Below $70,000
Investor Takeaway
Investors should be cautious of potential cybersecurity risks associated with AI tools.
More in General

Bitcoin Price Sinks 6% Below $66,500 Amid ETF Outflows and Institutional Selling

Bitcoin's Inflation-Hedging Potential Erodes as Price Falls Below $70,000

Bitcoin Falls Below $71,000 as Market Sees Widespread Liquidation and Heightened Geopolitical Risk
