CVE-2026-105697: Langflow arbitrary command execution through MCP stdio server configuration
- Source
- NVD(opens in a new tab)
- Identifier
- CVE-2026-105697
- Published
- Record updated
- Affected
- langflow >= 1.1.2, < 1.10.3, fixed in 1.10.3, released July 23, 2026, 74 days before disclosure
- langflow-base >= 0.1.2, < 0.10.3, fixed in 0.10.3, released July 23, 2026, 74 days before disclosure
- lfx < 1.10.3, fixed in 1.10.3, released July 23, 2026, 74 days before disclosure
- Known exploitation
- Not listed in the CISA Known Exploited Vulnerabilities catalog at the last check.
- EPSS
- 1.5%
Summary
CVE-2026-105697 affects Langflow before 1.10.3, where the MCP stdio transport ran whatever command and args a user placed in an MCP server configuration, with no allowlist and, before 1.10.3, wrapped in bash -c "exec {command} ...". Any user who can reach the MCP server settings (POST/PATCH /api/v2/mcp/servers/{server_name}) or build a flow with the MCP Tools component can run an arbitrary OS command on the Langflow host as the Langflow process user, when Langflow connects to the server, even if the UI then reports a startup failure. With the default LANGFLOW_AUTO_LOGIN=true, GET /api/v1/auto_login issues a token without credentials, so the flaw is reachable without an account on an exposed default instance; with AUTO_LOGIN disabled, any authenticated non-admin user can exploit it.
Mitigation
Fixed in Langflow 1.10.3, langflow-base 0.10.3, and lfx 1.10.3.
Affected packages in the Exposure Registry
Matched by package name and ecosystem. Each entry shows whether the package delegates to a language model and how many tracked packages depend on it.
- langflowPyPIReaches an LLM package through dependencies
- langflow-basePyPILLM dependency since 2024-04-01 · 1 tracked dependent
- lfxPyPILLM dependency since 2025-07-29 · 25 tracked dependents
Coverage
News, research and other items that mention CVE-2026-105697.
- Langflow's Critical Authenticated RCE Flaw Targeted in 1,423 Honeypot Exploit AttemptsThe Hacker News · 2026-10-15
Topics
Related items
- CriticalGHSA-9x8x-crjq-g64h: Jupyter MCP Server: Browser DNS rebinding can hijack the Jupyter backend through management routesSimilar attack · GitHub Advisory Database
- HighGHSA-24pw-rf5p-mgp7: Jupyter MCP Server: Unauthenticated management routes can hijack the Jupyter backend in standalone HTTP modeSimilar attack · GitHub Advisory Database
- HighGHSA-m932-63jc-4jqq: k8s-mcp-server: Missing authentication on default SSE/HTTP transport allows unauthenticated Kubernetes tool executionSimilar attack · GitHub Advisory Database
- HighGHSA-3mgh-93r5-r7jf: CodeWhale: exec_shell parallel-readonly classifier auto-approves fd/rg/git grep exec flags enabling silent RCESimilar attack · GitHub Advisory Database
- HighGHSA-qx49-fqc8-xw99: Model Context Protocol Python SDK: OAuth client could send credentials to an authorization server chosen by the MCP serverSimilar attack · GitHub Advisory Database