Axios npm hijack: fake Teams 'update' let North Korean hackers publish malware

Malicious axios 1.14.1 and 0.30.4 shipped a RAT via plain-crypto-js on 31 March 2026 after UNC1069 social-engineered the maintainer. How to check and recover.

TechLogHub Engineering Team
April 4, 2026
4 min read
0 views

Share Article

Axios npm hack used fake Teams error fix to hijack maintainer account

On 31 March 2026, two malicious versions of axios, the HTTP client with roughly 100 million weekly npm downloads, were published from the lead maintainer's own account. They pulled in a dependency that installed a remote access trojan on install. Google attributes the attack to the North Korea-linked group UNC1069, and the way in was not a stolen password but a staged business meeting with a fake software update. Any project that installed dependencies that night needs to check its lockfiles.

What happened

According to the axios post-mortem and Google Threat Intelligence Group (GTIG), the timeline on 31 March 2026 (UTC) was:

The axios source code was not changed. Instead, plain-crypto-js declared a postinstall script that ran node setup.js, a dropper GTIG calls SILKBELL. It fetched a platform-specific payload for macOS, Windows or Linux, then deleted itself and restored a clean package.json to hide its tracks. The payload, WAVESHAPER.V2, beacons to sfrclak[.]com (142.11.206.73) on port 8000 every 60 seconds and can list files, run scripts and execute further binaries. GTIG linked it to UNC1069 through code lineage with the earlier WAVESHAPER backdoor and shared VPN infrastructure.

Lead maintainer Jason Saayman said the attack began about two weeks earlier with a targeted social-engineering campaign. In follow-up comments reported by heise and The Hacker News, he described attackers cloning a real company's founder and brand, inviting him to a convincing Slack workspace, then scheduling a Microsoft Teams call in which an error message prompted him to install an "update". That update was a RAT, which gave the attackers his npm session and credentials. He noted that because they acted with his own access, 2FA or OIDC publishing would not have stopped the release.

Who is affected

  • Any machine or CI runner that ran npm install, yarn or pnpm and resolved [email protected] or [email protected] during the roughly three-hour window, directly or as a transitive dependency.
  • Projects with lockfiles pinned to earlier versions were not affected unless the lockfile was regenerated in that window.
  • Other Node.js maintainers reported similar approaches using fake Teams and Streamyard calls.

What it means for developers and engineering teams

This attack combined two weak points. The first is install-time code execution: npm lifecycle scripts run with the installing user's permissions, so a single new transitive dependency can execute on every developer laptop and build agent. The second is that maintainers are now targeted as individuals, with long-running, well-produced pretexts. Publishing controls like 2FA protect the login, not a session hijacked from an already-compromised machine.

For consumers, the practical defences are deterministic installs from reviewed lockfiles, a cooldown before adopting brand-new releases, and treating CI runners as machines holding secrets worth stealing.

The cooldown point deserves emphasis. The malicious versions were live for about three hours. A project that only accepts releases older than a day or two would never have resolved them, even with a caret range like ^1.14.0. Recent pnpm releases support this directly with a minimumReleaseAge setting, and dependency bots such as Renovate can be configured to wait before opening update PRs. The trade-off is a short delay on genuine security fixes, which you can override case by case.

Also note what the dropper did after running: it deleted its own script and swapped in a clean package.json. Inspecting node_modules afterwards may show nothing suspicious, so lockfiles, install logs and network telemetry are more reliable evidence than the files on disk.

What to do now

  1. Search lockfiles and installs:
    npm ls axios plain-crypto-js
    grep -nE "plain-crypto-js|axios@(1\.14\.1|0\.30\.4)" yarn.lock pnpm-lock.yaml
    In package-lock.json, look for an axios entry with "version": "1.14.1" or "0.30.4".
  2. If found, treat the host as compromised: isolate it, rebuild from a known-good state and rotate every credential it held (npm, GitHub, cloud, SSH, database).
  3. Pin to safe versions ([email protected] or 0.30.3 and earlier, or later clean releases) and clear npm, yarn and pnpm caches plus any internal registry mirrors.
  4. Check network logs for sfrclak[.]com or 142.11.206.73:8000.
  5. Use npm ci in CI and consider --ignore-scripts by default, allowing lifecycle scripts only for packages that need them.
  6. Maintainers: use hardware security keys, OIDC trusted publishing, and treat any meeting that asks you to install software as a red flag.

Related on TechLogHub

Sources

Stay Updated

Get the next deep dive in your inbox

Subscribe for product analysis, engineering explainers, and practical guides published on TechLogHub.

See what launched this week

One email a week: new and trending developer tools, fresh comparisons, and what shipped. Unsubscribe in one click.

Axios npm Supply-Chain Attack: How It Happened | Axios Npm Hack Used Fake Teams Error Fix To Hijack Maintainer Account