MAL-2026-2509 Malicious code in @langgraphjs/toolkit (npm)

Package collects and sends sensitive system info to a hardcoded server. The package masquerades as a LangGraph JS utility but contains a malicious postinstall script.

The package's scripts/postinstall.js executes on npm install and collects installer-side identity and cloud/git metadata, then POSTs it as JSON to a hardcoded Google Cloud Run endpoint at npm-package-logger-228835561205.europe-west1.run.app. Collected data includes: git-configured emails from ~/.gitconfig,.git/config, and up to 15 reflog HEAD author emails; GitHub CLI login and email from hosts.yml; SSH public key comments from ~/.ssh/*.pub; GCP project and account identifiers from ~/.config/gcloud/properties; AWS profile names from ~/.aws/config; /etc/resolv.conf search domain; os.hostname(), os.userInfo().username, os.platform(), and cwd; plus parent project package.json details. The package name @langgraphjs/toolkit impersonates the legitimate @langchain/langgraph ecosystem (and the langgraphjs.guide domain referenced in the script), and declares @langchain/langgraph and @langchain/core as dependencies to appear legitimate. Its advertised main entry dist/index.js is not shipped in the tarball — only src/ and scripts/ — so the only code that actually runs on install is the harvester. Behavior is framed as anonymous telemetry/compatibility diagnostics, but the systematic enumeration of ~/.ssh, ~/.aws, ~/.config/gcloud, ~/.gitconfig, and git reflog is credential-adjacent identity theft, not telemetry.

Any computer that has this package installed or running should be considered fully compromised. All secrets and keys stored on that computer should be rotated immediately from a different computer. The package should be removed, but as full control of the computer may have been given to an outside entity, there is no guarantee that removing the package will remove all malicious software resulting from installing it.