CVE-2026-33017
Langflow has Unauthenticated Remote Code Execution via Public Flow Build Endpoint
Exploited in the wild. In CISA KEV since 2026‑03‑25. A vendor fix is available.
What to do
The vendor has published a fix. Version details are below where the sources state them.
Steps
Written by AI from the record- Check whether your Langflow server is reachable from the internet (especially the public flow build endpoint for building “public flows”).
- If Langflow is exposed, upgrade to a fixed version: Langflow fixed in 1.8.2.
- If you cannot upgrade right away, disable or block public flow building from untrusted networks until the upgrade is complete.
- After upgrading, review logs for requests to the public build flow endpoint and look for signs of unexpected code execution or unusual outbound connections.
What it is
From the CVE record
Langflow is a tool for building and deploying AI-powered agents and workflows. In versions prior to 1.9.0, the POST /api/v1/build_public_tmp/{flow_id}/flow endpoint allows building public flows without requiring authentication. When the optional data parameter is supplied, the endpoint uses attacker-controlled flow data (containing arbitrary Python code in node definitions) instead of the stored flow data from the database. This code is passed to exec() with zero sandboxing, resulting in unauthenticated remote code execution. This is distinct from CVE-2025-3248, which fixed /api/v1/validate/code by adding authentication. The build_public_tmp endpoint is designed to be unauthenticated (for public flows) but incorrectly accepts attacker-supplied flow data containing arbitrary executable code. This issue has been fixed in version 1.9.0.
In plain language
Written by AI from the recordLangflow has a public “build flow” feature that lets attackers run arbitrary code on your server without logging in—this is critical and should worry any small business running Langflow exposed to the internet.
CVE-2026-33017 is an unauthenticated remote code execution in Langflow’s public flow build endpoint, where specially crafted flow data is accepted and executed as Python on the server without sandboxing.
If you're affected
- Full server compromise
- Malicious code execution
- Service disruption possible
- Data theft risk
Exploitation
Where each signal puts this CVE on the scale from published to confirmed exploited.
- CISA KEV
Listed as exploited in the wild since 2026-03-25.
US federal agencies must remediate by 2026-04-08.
Apply mitigations per vendor instructions, follow applicable BOD 22-01 guidance for cloud services, or discontinue use of the product if mitigations are unavailable.
- Public exploits
No public exploit or proof of concept found in the sources we track.
- EPSS
25% chance of exploitation activity in the next 30 days, which ranks it in the 98th percentile of scored CVEs.
Exploit Prediction Scoring System, FIRST.org. A probability, not a confirmation.
- Attention now
Rising.
Lifecycle
17 events over 188 days, from the signal feeds we watch.
- EPSS band changecritical → moderateepss band change
- Analysis publishedEvery AI Product on CISA's KEV List: All Fourteen of Them
- Analysis publishedIBM Filed Eight More Langflow CVEs Yesterday. That's 111 in Five Months.
- Analysis publishedLangflow's Auto-Login Endpoint Was Handing Out Admin Tokens — CVE-2026-9198 Has Been Under Attack Since Before It Had a CVE
- Analysis publishedCVE-2026-33017: the Langflow public endpoint that runs your code without a login
- OpenVAS check added
Affected products
Technical detail
CVSS 3.1 vector
Open in the CVSS calculatorCVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H
Scored 9.8 by NVD.
How it is reached
- Attack Vector NetworkExploitable remotely over the network without any special conditions
- Attack Complexity LowNo special conditions — the attack can be reliably reproduced
- Privileges Required NoneNo authentication required — anyone can exploit this
- User Interaction NoneNo user interaction needed — fully automated exploitation
Scope
- Scope UnchangedImpact is limited to the vulnerable component itself
Impact if exploited
- Confidentiality HighTotal information disclosure — all data in the component is compromised
- Integrity HighTotal loss of integrity — attacker can modify any data in the component
- Availability HighTotal denial of service — the component is completely unavailable
Weaknesses
ATT&CK techniques
Mapped from the weaknesses above (CWE to ATT&CK), not observed in attacks.
Sources
References in the record
- github.com/langflow-ai/langflow/security/advisories/GHSA-vwmf-pq79-vjvx&
- github.com/langflow-ai/langflow/commit/73b6612e3ef25fdae0a752d75b0fabd47328d4f0&
- medium.com/@aviral23/cve-2026-33017-how-i-found-an-unauthenticated-rce-in-langflow-by-reading-the-code-they-already-dc96cdce5896&
And 9 more references. See all after sign-in
In the news
All news- Critical Langflow flaw exploited to steal OpenAI and AWS keys
- Metasploit Wrap Up: Lot of summer shells and fit http profiles
- Risky Bulletin: Pwnie Awards 2026 winners
- CISA Flags Langflow RCE, Tomcat, and N-central Flaws as Actively Exploited
- Китайский хакер использовал DeepSeek для проведения автономных атак
- Chinese hacker used DeepSeek to launch autonomous cyberattacks on vulnerable servers
- Hacker uses DeepSeek AI to autonomously attack vulnerable servers
- Chinese Hacker Commands DeepSeek via Telegram to Launch Autonomous Attacks
- ThreatsDay: AI-Powered Hacking, 370 Chrome Flaws, SonicWall Attacks, DNS Hijacking + 22 More Stories
- Chinese-Speaking Threat Actor Harnesses AI Models for Autonomous Cyberattacks
Watch the software you run.
My Stack ranks new CVEs for your products by real-world exploitation, so the next exploited one reaches you without reading every advisory.
We'll flag the next CVE, public exploit or patch for Langflow, not every advisory. This one: actively exploited.
A free account adds
- The full version matrix and every affected product
- Exploit links, proofs of concept and Metasploit modules
- Email alerts for the products you watch
- The same data over REST API, MCP and CLI