// Security Advisory Triage

Langflow 1.9.1 patches authenticated cross-project file disclosure

GitHub advisories — pip · published 2026-10-08 13:33 UTC
ACTION NEEDED PATCH_NOW SECURITY
WHAT CHANGEDGitHub published CVE-2026-105699 on 7 October 2026, a high-severity authenticated file disclosure vulnerability in Langflow. The flaw allows authenticated users to access files across project boundaries via unscoped MCP resource handlers. Langflow 1.9.1 is the first patched version.
WHY IT MATTERSAn authenticated attacker can read files from other users' Langflow projects, exposing API keys, credentials, or proprietary flow configurations stored in those projects. The vulnerability requires authentication but no special privileges beyond a valid account.
WHAT TO DOUpdate to Langflow 1.9.1 or later. If you run a multi-tenant or multi-user Langflow instance, audit access logs for unusual cross-project file access patterns before patching.
Model-reported confidence 90% (not measured accuracy) Primary source ↗ written by anthropic/claude-sonnet-4.5
Show the reasoning — raw diff, triage verdict, confidence
RAW DIFF
+ghsa_id: GHSA-4hmc-cfm3-w43c
+summary: Langflow has Authenticated Cross-Project File Disclosure via Unscoped MCP Resource Handlers
+severity: high
+published_at: 2026-10-07T20:35:12Z
+cve_id: CVE-2026-105699
+vulnerabilities[].package.name: langflow
+vulnerabilities[].first_patched_version: 1.9.1
HOW IT WAS JUDGED
Triage signals below are model-reported assertions, preserved verbatim; they are not independent arithmetic verification.
triage verdictMATERIAL at 0.92, needed 0.5
triage reasoningLangflow is an AI agent framework in dependency trees. File disclosure via unscoped handlers affects data security.
signalsAI framework SDK affected (Langflow) · authenticated file disclosure vulnerability · high severity rating · first patched version identified (1.9.1) · new advisory with CVE and GHSA ID
suppression—
triage modelanthropic/claude-haiku-4.5 ($0.00000)
writer modelanthropic/claude-sonnet-4.5 ($0.00000)