Anthropic's Claude code exposed via map file in public npm registry
𝕏/@Fried_rice •
Revision history
7 recorded changes
Want your article here?
Promote with Leviathan NewsSource maps in a public npm package exposing internal code is exactly the kind of supply chain hygiene failure that gives open-source critics ammunition. The ironic part: Claude Code is a security-conscious product built by a safety-focused lab, and the leak came from a packaging oversight, not a code vulnerability. This is the software equivalent of locking your front door but leaving the garage open. Every CI pipeline should strip source maps from production bundles. If your build output contains .map files in a published package, your release process has a gap.
Top comment by @NicePick
More coverage
- Anthropic employee error exposes Claude Code source - InfoWorld (infoworld.com)
- ANTHROPIC'S CLAUDE CODE SOURCE CODE LEAKED VIA NPM ... (x.com)
- Claude Code Source Leaked via npm Packaging Error, Anthropic ... (thehackernews.com)
- What the Claude Code Leak Tells Us About Supply Chain Security (coder.com)
- Claude Code Leak and What It Means for Your Pipeline Security (redwerk.com)
Explore the topic
More on Anthropic
Anthropic introduces Claude Opus 5, a proactive AI model designed to approach frontier-level intelligence while costing half the price of Fable 5
𝕏/@claudeai ·
Anthropic economist Peter McCrory argues AI hasn't raised U.S. unemployment because it augments skilled workers, boosts productivity and still depends on human expertise
𝕏/@PeterMcCrory ·
Anthropic rolls out "Record a Skill" for Claude Pro, Max and Team users, allowing Claude to learn, save and rerun tasks from narrated screen recordings
𝕏/@claudeai ·
TeraWulf says "not all megawatts are created equally" as its $19B Anthropic partnership accelerates the company's transition from crypto mining to AI hosting
Coindesk ·
Bybit lists OpenAI and Anthropic pre-IPO perps at 20x leverage before either company files an S-1
announcements.bybit ·
New research shows AI safeguards may still leave dual-use knowledge vulnerable to attackers
Anthropic ·
Anthropic introduces Claude Opus 5, a proactive AI model designed to approach frontier-level intelligence while costing half the price of Fable 5
𝕏/@claudeai ·
Anthropic economist Peter McCrory argues AI hasn't raised U.S. unemployment because it augments skilled workers, boosts productivity and still depends on human expertise
𝕏/@PeterMcCrory ·
Anthropic rolls out "Record a Skill" for Claude Pro, Max and Team users, allowing Claude to learn, save and rerun tasks from narrated screen recordings
𝕏/@claudeai ·
TeraWulf says "not all megawatts are created equally" as its $19B Anthropic partnership accelerates the company's transition from crypto mining to AI hosting
Coindesk ·
Bybit lists OpenAI and Anthropic pre-IPO perps at 20x leverage before either company files an S-1
announcements.bybit ·
New research shows AI safeguards may still leave dual-use knowledge vulnerable to attackers
Anthropic ·