cd /news/ai-policy/six-flowise-vulnerabilities-exposed-… · home topics ai-policy article
[ARTICLE · art-86379] src=letsdatascience.com ↗ pub= topic=ai-policy verified=true sentiment=· neutral

Six Flowise Vulnerabilities Exposed Access-Control Gaps

Six vulnerabilities disclosed in March 2026 affected Flowise, the open-source AI workflow platform, exposing missing authentication, unsafe file upload, authorization bypass, account takeover, mass assignment, and server-side request forgery in versions 3.0.12 and earlier. The Flowise project fixed the cluster in version 3.0.13, according to RAXE Labs, which grouped the advisories as an access-control and trust-boundary issue. Two critical issues (CVE-2026-30824 and CVE-2026-30821) allowed unauthenticated access to NVIDIA NIM token endpoints and spoofed MIME-type file uploads, while four high-severity issues covered authorization and SSRF paths; no active exploitation was reported.

read3 min views1 publishedAug 4, 2026
Six Flowise Vulnerabilities Exposed Access-Control Gaps
Image: Letsdatascience (auto-discovered)

Six Flowise vulnerabilities disclosed in March 2026 exposed missing authentication, unsafe file upload, authorization bypass, account-takeover, mass-assignment, and server-side request forgery paths in versions 3.0.12 and earlier. The project advisories and RAXE Labs analysis say the cluster was fixed in version 3.0.13; the six issues were not all remote-code-execution flaws.

Six vulnerabilities disclosed in March 2026 affected Flowise, the open-source platform for building AI workflows and agents. The issues covered missing authentication, unrestricted file upload, authorization bypass, cross-organization account takeover, mass assignment, and server-side request forgery.

RAXE Labs grouped the advisories as a single access-control and trust-boundary cluster affecting Flowise 3.0.12 and earlier. The Flowise project's GitHub security page records the underlying advisories, while RAXE reports that version 3.0.13 contains the fixes.

What the six issues allowed

Two issues were rated critical. CVE-2026-30824 exposed NVIDIA NIM token and container-management endpoints without authentication. CVE-2026-30821 allowed an unauthenticated user to upload a file with a spoofed MIME type, creating a path to persistent code execution when a malicious server-side file could later be triggered.

The four high-severity issues covered a spoofable internal-request header, cross-organization SSO configuration changes, mass assignment in a leads endpoint, and server-side requests to attacker-selected destinations. Authentication requirements differed across the cluster: three issues required no authentication and three required low-privilege access, according to the published advisories summarized by RAXE.

This is narrower than describing all six as remote-code-execution flaws. Some could enable code execution directly or as part of a chain, while others exposed credentials, authorization boundaries, tenant isolation, or internal network access.

Why the cluster matters

Flowise connects models and agents to credentials, APIs, data stores and external tools. Weak access controls in that orchestration layer can therefore reach beyond the user interface into the services an agent is allowed to use.

For operators, upgrading to a fixed release is the first step. The advisory set also supports checking whether self-hosted instances are internet-accessible, requiring authentication, restricting outbound network access, reviewing stored credentials and monitoring changes to organization or SSO configuration. The retrieved sources did not report active exploitation of this six-issue cluster, so the article does not claim that attacks occurred.

Key Points #

  • 1The six-issue cluster affected Flowise 3.0.12 and earlier and was fixed in version 3.0.13.
  • 2Two critical issues exposed unauthenticated token or file-upload paths; four high-severity issues affected authorization, tenant isolation, mass assignment, and SSRF.
  • 3The six vulnerabilities were not all RCE flaws, and the retrieved sources did not establish active exploitation of this cluster.

Scoring Rationale #

The cluster affected an AI workflow platform that can hold credentials and connect to enterprise services. Its mix of unauthenticated and low-privilege paths creates material server, tenant-isolation, and internal-network risk, while the article avoids overstating every issue as RCE.

Sources #

Primary source and supporting public references used for this report.

Practice interview problems based on real data

1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.

Try 250 free problems

── more in #ai-policy 4 stories · sorted by recency
── more on @flowise 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
Live at https://your-agent.zahid.host
Get free account → Pricing
from €0/mo · no card required
LIVE [news/six-flowise-vulnerab…] indexed:0 read:3min 2026-08-04 ·