Security Digest · Vulnerability record
CVE-2026-105697: Langflow: OS command injection (RCE) via arbitrary command in MCP stdio server configuration
Langflow is a tool for building and deploying AI-powered agents and workflows.
CVSS 9.9 Critical EPSS 0.5% · p40
Should you act?
Critical, preconditions apply
Severe if exploited, but exploitation needs conditions that may not apply to you. This signal is computed from CISA KEV, EPSS and CVSS data; it is not an editorial verdict.
- No confirmed exploitation. It is not in CISA’s KEV catalog and CISA’s SSVC assessment records no known exploitation. That can change quickly after a public write-up; re-check if you are affected.
- EPSS 0.5%, percentile 40: FIRST’s model estimates a 0.5% chance of exploitation activity in the next 30 days, higher than 40% of all scored CVEs (low). EPSS is a probability, not a verdict.
- Needs an account (low (any account)). Exposure depends on who can log in.
- Only applies if your dependency tree includes
langflow(PyPI) >= 1.1.2, < 1.10.3, fixed in 1.10.3;langflow-base(PyPI) >= 0.1.2, < 0.10.3, fixed in 0.10.3;lfx(PyPI) < 1.10.3, fixed in 1.10.3. Check your lockfile, including transitive dependencies.
What it is
Langflow is a tool for building and deploying AI-powered agents and workflows. Before Langflow 1.10.3, the MCP stdio transport launched whatever command / args a user put in an MCP server configuration, with no allowlist and (before 1.10.3) wrapped in bash -c "exec {command} ...". Any user able to reach the MCP server settings ("Settings → MCP Servers → Add MCP Server", POST/PATCH /api/v2/mcp/servers/{servername}) or to build a flow with the MCP Tools component could add a "server" whose command is an arbitrary OS command (touch, rm -rf, a reverse shell, ...). The command runs on the Langflow host as the Langflow process user as soon as Langflow tries to connect to the server (listing servers, loading tools, running the flow) — even when the UI then reports that the stdio server failed to start. With the default LANGFLOWAUTOLOGIN=true, GET /api/v1/autologin hands out a token without credentials, so on an exposed instance running the default configuration this is reachable without an account. AUTOLOGIN is documented as a development-only setting; with it disabled, any authenticated (non-admin) user can exploit it. This issue is fixed in Langflow 1.10.3, langflow-base 0.10.3, and lfx 1.10.3.
From the GitHub advisory
Summary
Before Langflow 1.10.3, the MCP stdio transport launched whatever command / args a user put in an MCP server configuration, with no allowlist and (before 1.10.3) wrapped in bash -c "exec {command} ...". Any user able to reach the MCP server settings ("Settings → MCP Servers → Add MCP Server", POST/PATCH /api/v2/mcp/servers/{servername}) or to build a flow with the MCP Tools component could add a "server" whose command is an arbitrary OS command (touch, rm -rf, a reverse shell, ...). The command runs on the Langflow host as the Langflow process user as soon as Langflow tries to connect to the server (listing servers, loading tools, running the flow) — even when the UI then reports that the stdio server failed to start.
With the default LANGFLOWAUTOLOGIN=true, GET /api/v1/autologin hands out a token without credentials, so on an exposed instance running the default configuration this is reachable without an account. AUTOLOGIN is documented as a development-only setting; with it disabled, any authenticated (non-admin) user can exploit it.
The issue was fixed in two steps and is fully fixed in Langflow 1.10.3 (and 1.11.0+):
- 1.9.0 — #12290 added a command allowlist and argument/env validation to the REST model (MCPServerConfig), blocking the "Add MCP Server" vector reported here. - 1.10.3 — #14036 applied the same policy at the execution sink (lfx.base.mcp.util), including configs embedded in flows / tweaks and the final pre-spawn boundary, and removed the bash -c wrapper (the process is now exec'd directly, without a shell).
Weakness type: CWE-78.
Product: langflow (PyPI) · Assigned by: GitHub_M · Published: 2026-10-05
Affected and fixed versions
| Ecosystem | Package | Vulnerable | First fixed |
|---|---|---|---|
| PyPI | langflow | >= 1.1.2, < 1.10.3 | 1.10.3 |
| PyPI | langflow-base | >= 0.1.2, < 0.10.3 | 0.10.3 |
| PyPI | lfx | < 1.10.3 | 1.10.3 |
| Vendor | Product | Affected versions (per CVE record) |
|---|---|---|
| langflow-ai | langflow | >= 1.1.2, < 1.10.3 |
| langflow-ai | langflow-base | >= 0.1.2, < 0.10.3 |
| langflow-ai | lfx | < 1.10.3 |
Severity scores
| Scored by | CVSS | Score | Vector |
|---|---|---|---|
| GitHub_M | 3.1 | 9.9 Critical | CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:C/C:H/I:H/A:H |
| GitHub | 3.1 | 9.9 | CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:C/C:H/I:H/A:H |
- Attack vector: network (remotely reachable)
- Attack complexity: low
- Privileges required: low (any account)
- User interaction: none
- Scope: changed (can affect other components)
CVSS measures how bad a flaw is if exploited, not how likely exploitation is. Read it next to the exploitation data below.
Exploitation and urgency
- CISA KEV: not listed (no confirmed in-the-wild exploitation known to CISA).
- CISA SSVC assessment: exploitation none, automatable no, technical impact total.
- FIRST EPSS (model of 2026-10-21): 0.47% probability of exploitation activity in the next 30 days; percentile 40.
Timeline
- — CVE record published by GitHub_M
- — GitHub-reviewed advisory GHSA-w794-rj3p-xv45 published (critical)
- — CVE record last updated
Sources and references
Original sources first. Version and fix details should always be confirmed against the vendor or project advisory.
Vulnerability records
- GitHub Advisory GHSA-w794-rj3p-xv45 github.com
- CVE.org record CVE-2026-105697 cve.org
- NVD entry CVE-2026-105697 nvd.nist.gov
Other references from the CVE record and advisory
- github.com github.com
- github.com github.com
- github.com github.com
- github.com github.com
- github.com github.com
- github.com github.com
- github.com github.com
- ox.security ox.security
- ox.security ox.security