Weaponizing trust signals: Claude Code lures and GitHub release payloads

Summary: Trend Micro reveals how an accidental exposure of internal Claude Code code on npm allowed threat actors to quickly distribute malware through false GitHub repositories.

Trend Micro's investigation starkly illustrates how attackers convert a data breach into operational campaigns. Following the accidental publication of internal code linked to Claude Code, malicious actors capitalized on the generated attention by setting up fake repositories on GitHub and distributing malware masquerading as supposed leaks or recovered ‘builds’.

The most unsettling detail is not just the initial leak but how it was exploited to manipulate trust. The attackers used signals that still function as legitimacy shortcuts for many users: brand name, GitHub Releases, large files, a real software appearance, and disposable accounts that allow them to reappear after each takedown.

The story is significant because it reveals a mutation of classical deception in developer environments. It’s no longer enough to impersonate a web page or email; now attackers are mimicking distribution and collaboration flows that are part of modern software routines. When the bait appears as a plausible technical download, the psychological barrier drops much faster.

At its core, this case is not just about Claude Code. It is about how threat actors learn to monetize attention, urgency, and trust in legitimate platforms within hours.

Key facts

  • Anthropic exposed approximately 512,000 lines of internal TypeScript code in its npm package @anthropic-ai/claude-code.
  • Threat actors created false GitHub repositories to distribute malware that was disguised as 'leaked' or recovered ‘builds’ of Claude Code.
  • The accidental exposure led to a campaign that included the use of GitHub Releases as a reliable channel for delivering malware.

Why it matters

This incident underscores the importance of constant monitoring in the software lifecycle and the risks associated with human and organizational vulnerabilities, demonstrating how a simple error can be exploited to cause significant damage.

X profile@trendaisecurityhttps://x.com/trendaisecurity
Embedded content for: Weaponizing trust signals: Claude Code lures and GitHub release payloads