Breaking

Authenticated-free RCE in Langflow exploited to mine Monero within 20 hours

June 30, 2026 at 11:54 EDT

Security
AI Agents
Open Source

An unauthenticated remote code execution (RCE) flaw in the open-source AI workflow builder Langflow, tracked as CVE-2026-33017, is being actively exploited to run arbitrary Python code and deploy a Monero miner, according to analyses by multiple security firms. The U.S. CISA has added the vulnerability to its Known Exploited Vulnerabilities (KEV) catalog and is urging immediate patching.

March 17, 2026 · CVE-2026-33017

One HTTP request to Monero: critical Langflow flaw exploited in the wild

An authentication-free API endpoint in the popular AI workflow builder Langflow lets attackers run arbitrary Python on the server — and within 20 hours of disclosure, it was being weaponized to deploy crypto-mining malware.

9.8

CVSS score — rated Critical

Single HTTP request for full RCE

~145k

GitHub stars — broad exposure

Speed of exploitation

Time from public disclosure to observed in-the-wild attacks.

20 hrs

Disclosure →
exploitation

24 hrs

One full
day

Attackers struck in less than a day — defenders had almost no window to patch.

The attack chain

Auth-free endpoint

POST /build_public_tmp — accepts attacker flow JSON

→

exec() runs code

Arbitrary Python passed unsandboxed

→

Lambsys dropped

Go malware, kills rivals, persists, hides logs

→

Monero mining

XMR mined; lateral move via SSH keys

Are you affected?

A deliberate design choice — handling public flows without authentication — opened the door to unintended code execution.

Vulnerable

v1.8.2 and earlier

Fixed

v1.9.0 and later

Interim measures: avoid direct internet exposure · strengthen WAF and authentication · patch promptly — now on the CISA Known Exploited Vulnerabilities catalog.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.