Claude Code Source Code Leaked, Over 500,000 Lines of Code Obtained by Developers

marsbitPublicado a 2026-04-01Actualizado a 2026-04-01

Resumen

Anthropic faces a major security incident as the complete source code for its AI tool Claude Code, exceeding 500,000 lines of TypeScript, was leaked. The breach originated when a developer shared a compressed file on Twitter, which quickly gained over 5.3 million views. The exposure was due to Anthropic accidentally including .map files when publishing to npm, revealing core code. The leaked code unveils several advanced features, including a project codenamed BUDDY, which offers a personalized cyber pet system with traits like debugging power and even a "sarcasm value." Another feature, KAIROS, aims to create an "always-on" AI assistant capable of maintaining memory across sessions and initiating tasks. The code also includes a "Nightly Dreaming" mechanism, allowing Claude to organize information and consolidate core content while the user sleeps. Despite the breach, the incident has inadvertently showcased Anthropic's innovative capabilities, turning a PR crisis into an unexpected demonstration of their cutting-edge AI developments.

Recently, artificial intelligence company Anthropic encountered an unprecedented crisis when the complete source code of its latest tool, Claude Code, was accidentally leaked within the developer community. The incident began when a developer, Chaofan Shou, posted a compressed package containing over 500,000 lines of TypeScript code on Twitter, sparking widespread attention and discussion among developers globally. This leak quickly exploded online, with page views surpassing 5.3 million within hours, becoming a hot topic in the tech circle.

So, what was the cause of this leak? According to reports, when Anthropic published the code to npm, it actually forgot to remove the .map files, leading to the exposure of the core code. This mistake not only embarrassed the company but also raised external doubts about its security management. However, this accidental leak has given the outside world an early glimpse into some of Anthropic's cutting-edge functionalities in the AI field.

From the source code, it can be seen that Claude Code includes several astonishing features, with the project codenamed BUDDY being particularly notable. It can not only accompany developers in writing code but also has a unique personalized pet system, allowing users to incubate their own cyber pets based on personal information. This pet not only has a cute appearance but also has dynamic stats like debugging power and patience, and even a "sarcasm value," undoubtedly adding a lot of fun to the work atmosphere.

Even more forward-looking is the feature codenamed KAIROS, which aims to achieve an "always-on" AI assistant. Unlike traditional conversation methods, KAIROS can maintain memory across multiple sessions and proactively initiate tasks. Additionally, the source code includes a Nightly Dreaming mechanism, where Claude organizes information and consolidates core content while the user sleeps. This innovative concept not only makes the AI smarter but also brings new possibilities for future work methods.

Although Anthropic is currently busy dealing with this leak, the unexpected exposure might also provide a platform to showcase their cutting-edge technology. This accidental PR crisis has instead highlighted the company's innovative capabilities. The future of the AI field will become even more exciting because of such attempts.

Preguntas relacionadas

QWhat was the cause of the Claude Code source code leak at Anthropic?

AAnthropic forgot to remove .map files when publishing the code to npm, which exposed the core source code.

QHow many lines of TypeScript code were leaked in the Claude Code incident?

AOver five hundred thousand lines of TypeScript code were leaked.

QWhat is the name of the project that features a personalized cyber pet system for developers?

AThe project is called BUDDY, which includes a personalized cyber pet system with traits like debugging power and patience.

QWhat is the KAIROS feature designed to do according to the leaked source code?

AKAIROS is designed to be an 'always-on' AI assistant that maintains memory across multiple sessions and can proactively initiate tasks.

QWhat mechanism does Claude have for processing information during user inactivity?

AClaude has a 'Nightly Dreaming' mechanism where it organizes information and consolidates core content while the user is sleeping.

Lecturas Relacionadas

24/7 Unstoppable Derivatives Wave: Cryptocurrency Is Forcing Traditional Finance to 'Change Time Zones'

The article discusses how the 24/7 nature of the cryptocurrency market is compelling traditional finance to adapt its operating hours and infrastructure. The key catalyst is the CME Group's planned launch of nearly round-the-clock trading for regulated crypto derivatives, a move driven by strong institutional demand for continuous risk management. This shift highlights a fundamental change: derivatives, not spot trading, now dominate crypto market activity and price discovery. However, integrating continuous trading into traditional finance reveals structural tensions. While execution times can be extended, settlement, clearing, and regulatory reporting largely remain bound to traditional business-day cycles. This creates a lag where weekend price movements can impact risk exposures before traditional control systems are fully active. Furthermore, the article explores new challenges arising from this always-on environment. The inherent transparency of public blockchains, while ensuring auditable settlement, also exposes sensitive corporate information like treasury flows to competitors in real-time. This has elevated privacy from a feature to a core requirement for institutional adoption. The next phase hinges on building systems that balance this necessary privacy with regulatory accountability and compliance. In conclusion, the move towards 24/7 trading signifies more than crypto becoming institutionalized. It represents traditional finance beginning to adopt the temporal structure of crypto-native markets. The future will be defined by how successfully traditional risk, identity, privacy, and settlement frameworks can operate at the continuous speed cryptocurrency markets demand.

marsbitHace 32 min(s)

24/7 Unstoppable Derivatives Wave: Cryptocurrency Is Forcing Traditional Finance to 'Change Time Zones'

marsbitHace 32 min(s)

From "Selling Cloud" to "Selling Token": Telecom Operators Fully Join the AI Battlefield

From selling cloud services to tokens, operators fully join the AI battlefield. Major Chinese telecom operators—China Telecom, China Mobile, and China Unicom—are launching AI token packages for personal, family, and enterprise use. This marks a shift from selling traditional services like voice and data to offering tokens as a new standard unit for AI service consumption. Operators are transforming into AI service integrators, moving beyond their role as cloud providers. They are building an integrated structure that combines models, computing power, applications, unified billing, and security. By bundling various AI models, pooling heterogeneous computing resources, developing industry-specific agents, and leveraging their existing billing systems, they aim to lower entry barriers and simplify AI adoption for businesses. The token model is becoming a key packaging method for AI services, moving beyond developer markets to the mainstream. Operators are positioning themselves as the essential infrastructure for the AI era, aiming to make AI a billable, manageable, and reliable utility similar to electricity or network services. However, challenges remain in ensuring model quality, cost reduction, and achieving sustainable commercial success. The ultimate goal is to make AI a measurable and scalable production factor for industries.

marsbitHace 54 min(s)

From "Selling Cloud" to "Selling Token": Telecom Operators Fully Join the AI Battlefield

marsbitHace 54 min(s)

Unitree Passes the Hearing, Hangzhou Reaps the Rewards

Unitree Technology, a leading company in Hangzhou's tech scene known as one of the "Hangzhou Six Dragons," has officially passed the review for listing on the Shanghai Stock Exchange's STAR Market (科创板). It plans to raise 4.202 billion yuan for the research and development of intelligent robot models and robot hardware. This milestone will make Unitree the "first humanoid robotics stock." Founded in 2016 by Wang Xingxing, the company started humbly in a small office in Hangzhou's Binjiang district. Initially, the robotics sector was not viewed favorably by the market, with Unitree's products often labeled as "toys" and struggling to secure funding. At its most critical point, with only around 100,000 yuan left, Wang stopped his own salary to keep the company afloat. A crucial turning point came in 2018 when Hangzhou's state-owned capital system provided timely support. A financial platform under the city's state-owned assets completed due diligence in three days and granted a 20-million-yuan loan within a week. This "patient capital" infusion stabilized Unitree, enabling its transition from prototype development to mass production and commercial viability. Subsequently, Hangzhou Capital, through its two major 100-billion-yuan mother funds—the Hangzhou Science and Technology Innovation Fund and the Hangzhou Innovation Fund—participated in four of Unitree's financing rounds (B2, B3, C, and C+). This continuous backing helped the company grow, attract top-tier industrial investors like China Mobile, Tencent, Alibaba, and Geely, and solidify its position as a global leader in legged robotics. By 2025, Unitree achieved significant scale, with revenue reaching 16.99 billion yuan, net profit of 5.91 billion yuan, global leadership in humanoid robot shipments, and over 33,000 quadruped robots sold worldwide. Unitree's journey exemplifies Hangzhou's strategy of nurturing hard-tech startups from "seedlings" to industry leaders. Beyond Unitree, Hangzhou's capital ecosystem has supported other "Six Dragons" like Cloudwalk, BrainCo, and DeepSeek. The city has established a 500-billion-yuan "3+N" industrial fund cluster and specialized early-stage funds like the "Runmiao Fund" with a 20-year term to fill funding gaps for very early-stage projects. This robust "capital + talent" model, coupled with an influx of over 430,000 young professionals in 2025 alone, has fostered a vibrant innovation ecosystem. Hangzhou is now home to 48 unicorns and 413 potential unicorns, building comprehensive industrial chains in AI, robotics, brain-computer interfaces, and more. As Hangzhou experiences a wave of IPOs, it is solidifying its reputation as an ideal city for entrepreneurs.

marsbitHace 57 min(s)

Unitree Passes the Hearing, Hangzhou Reaps the Rewards

marsbitHace 57 min(s)

Mystery Trader Dumps $1.26B in BlackRock Bitcoin ETF

A mystery trader sold a massive block of 29.2 million shares of BlackRock's iShares Bitcoin Trust (IBIT) for $1.26 billion via a dark pool. NYDIG's Greg Cipolaro analyzed the trade, noting the seller accepted a price $1.01 below market value, forfeiting $29.5 million for immediate execution. He speculates this indicates a large directional holder exiting a concentrated position, not a basis-trade unwind. The sale occurred amid a broader trend of outflows from U.S. spot Bitcoin ETFs, which have seen 11 consecutive days of net withdrawals totaling over $2.9 billion since mid-May. Bitcoin's price dipped 2.8% following the block trade, but analysts noted the market absorbed the large sale relatively well. Overall market sentiment remains cautious, with a key "fear" index averaging in "fear" territory for May.

TheNewsCryptoHace 1 hora(s)

Mystery Trader Dumps $1.26B in BlackRock Bitcoin ETF

TheNewsCryptoHace 1 hora(s)

Are Rising U.S. Stocks Getting More Dangerous? Goldman Sachs: Downside Protection Mechanisms Have Almost Failed

The US stock market rally is showing signs of becoming increasingly precarious as key downside protection mechanisms fail, according to Goldman Sachs. Derivatives strategist Brian Garrett notes that the S&P 500 options volatility skew has plunged to an 18-month low, indicating the market now prices an 8% probability for both a 10% drop and a 10% rise—a sign of "skew failure." Concurrently, Goldman's Panic Index hit a two-year low, reflecting minimal demand for tail-risk hedging. This complacency emerges amid a relentless market surge, with the S&P 500 setting new records frequently in 2024. Garrett highlights three major concerns: extreme concentration in the top ten stocks (40% of index weight), heavy reliance on AI-themed performance, and a price pattern eerily similar to the 1998-1999 period. Despite pervasive media pessimism, this fear is absent in options pricing. Downside hedge costs are historically low. Goldman suggests tactical trades: buying RSP outperformance options versus the SPX for a broadening rally, purchasing VIX calls for protection, and going long on Bitcoin ETF volatility. Hedge funds have been net buyers for two weeks, with sector rotation into financials and out of industrials. Notably, the global single-stock leveraged/ inverse ETF AUM has doubled to over $60 billion in two months, underscoring growing speculative activity.

marsbitHace 1 hora(s)

Are Rising U.S. Stocks Getting More Dangerous? Goldman Sachs: Downside Protection Mechanisms Have Almost Failed

marsbitHace 1 hora(s)

Trading

Spot

Futuros