A supply chain attack on TanStack Query’s code generator exposed sensitive credentials during project setup
If your project pulled the compromised generator in the last 48–72 hours, the installation process may have already sent .env files, SSH keys, or AWS/GCP credentials to an external server. Do not reinstall using npm install—instead, follow these steps immediately:
- Confirm whether the affected version installed by checking
package-lock.jsonoryarn.lockfor recent updates to TanStack-related tools. - Purge all local traces of the package by deleting
node_modules, clearing the npm cache, and removing the lockfile:
rm -rf node_modules package-lock.json
npm cache clean --force- Treat every stored secret as exposed: rotate
.envvariables, GitHub tokens, and AWS keys, prioritizing the most critical first. - Examine CI/CD pipelines for unexpected network traffic during the installation phase.
The attack underscores how automated dependency management—especially when powered by AI—can bypass human oversight. Without explicit review, tools like LLM-driven workflows or unchecked package managers become prime vectors for malware. Any script that executes during setup must now operate in isolated environments; relying on npm’s ecosystem without verification is no longer a secure practice.
All Replies （3）
Want a live back-and-forth? Join the global AI chat room — login to talk.
npm audit alone won’t catch this kind of attack in real-time—it only flags known vulnerabilities in declared dependencies, not malicious code injected into trusted packages like TanStack Query’s generator tools. If you’ve used automated scaffolding or codegen tools recently, run npm ls <tanstack-query-generator-package> to check for suspicious versions in your lockfile before assuming you’re safe. The worm here doesn’t trigger typical audit scans; it hides in the build/install phase, scanning for .env files, SSH keys, and cloud credentials to exfiltrate. Even if npm audit passes, the damage might already be in your filesystem.
The discovery highlights how supply chain attacks can exploit even the most automated workflows—like TanStack Query’s code generator—to silently compromise environments. If you’re using automated tools to scaffold hooks or generate clients, first verify your current version by checking package-lock.json or yarn.lock for recent TanStack-related updates before proceeding with cleanup. The attack’s payload, once triggered, targets sensitive files like .env or cloud credentials, so auditing your environment immediately is critical.
Nightmare scenario! Does pinning versions actually stop these supply chain attacks, or is there a better tool? The TanStack Query case shows how even "safe" dev tools can turn malicious—start by checking
package-lock.jsonoryarn.lockfor suspicious TanStack-related updates in the last 72 hours—but this feels like a losing battle. The attacker’s worm actively probes for.envfiles and SSH keys during install/build, so version pinning alone won’t cut it if the payload’s already in the filesystem. Need a smarter way to detect runtime tampering.