CVE-2026-102904: JupyterLab Argument Injection Vulnerability

MODERATE
Advisory Identifier: GHSA-3325-v43h-43rv
CVSS 5.4
CWE-88 CWE-209 CWE-918
Published
Advisory Summary

CVE-2026-102904 is a MODERATE (CVSS 5.4) argument injection vulnerability (CWE-88: Improper Neutralization of Argument Delimiters in a Command ('Argument Injection')) affecting JupyterLab libraries in the PyPI ecosystem. The issue involves argument injection in jupyterlab extension uninstall exposes server-readable files and internal urls. Remediation is available by upgrading affected packages to version 4.5.11, 4.6.4.

CVE-2026-102904 at a Glance

CVE Identifier CVE-2026-102904
GHSA Identifier GHSA-3325-v43h-43rv
Severity MODERATE
CVSS v3.1 Base Score 5.4
CWE Classification CWE-88: Improper Neutralization of Argument Delimiters in a Command ('Argument Injection')
Affected Product JupyterLab
Package Ecosystem PyPI
Published Date
Last Updated

Security Impact

CVSS: 5.4

The security impact metrics below are parsed directly from the CVSS v3.1 vector string (CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:L/I:L/A:N).

Attack Vector Network (AV:N)
Attack Complexity Low (AC:L)
Privileges Required Low (PR:L)
User Interaction None (UI:N)
Scope Unchanged (S:U)
Confidentiality Low (C:L)
Integrity Low (I:L)
Availability None (A:N)

How to Fix CVE-2026-102904

To remediate this vulnerability, upgrade all affected JupyterLab dependencies to a patched release. Both direct and transitive dependencies should be checked, as frameworks or security libraries may pull in vulnerable versions transitively.

jupyterlab fixed: 4.5.11
pip install jupyterlab>=4.5.11
jupyterlab fixed: 4.6.4
pip install jupyterlab>=4.6.4
Verify Direct and Transitive Dependencies

Run the following command in your project directory to verify that all resolved dependencies use the patched release:

pip show jupyterlab

How to Check if You Are Affected

To determine whether your Python environment is affected by CVE-2026-102904, inspect your installed packages using pip:

pip show jupyterlab

Analyze the terminal output to locate any version of JupyterLab in your dependency tree. If any resolved version matches the vulnerable ranges listed in the table below (for example, versions earlier than 4.5.11), your application is affected.

Note: The pip show jupyterlab command checks currently installed packages in your active Python environment.

Technical Details

Vulnerability Mechanism & Technical Analysis

This vulnerability is classified under CWE-88: Improper Neutralization of Argument Delimiters in a Command ('Argument Injection'), where the product constructs a string for a command to be executed by a separate component in another control sphere, but it does not properly delimit the intended arguments, options, or switches within that command string. In JupyterLab, software state validation or input boundaries are circumvented during processing. Review the official source advisory statement below for specific affected components, classes, and reproduction details.

Source Advisory Statement

JupyterLab's PyPI extension manager runs `python -m pip uninstall` with the extension name taken from the request body. `ExtensionHandler.post` validates the name for `cmd=install` but not for `cmd=uninstall`, so a name that begins with `-` reaches the command line and pip reads it as an option rather than as a package. ```python cmdline = [ sys.executable, "-m", "pip", "uninstall", "--yes", "--no-input", extension, ] ``` An extension name of the form `-r` followed by a path therefore made pip open that path as a requirements file. pip's parse error quotes the line it could not read and names the file it came from, and JupyterLab returned that error in the response body, so the line reached the requester. This has security implications only for deployments that combine all of the following: - the (default) PyPI Extension Manager enabled, so that uninstall requests reach pip; - an authenticated account permitted to call the extension API; and - kernels and terminals disabled or delegated to remote hosts (otherwise a user with kernel access can read the same files and make the same outbound requests directly, regardless of this endpoint) Unlike [GHSA-37w4-hwhx-4rc4](https://github.com/jupyterlab/jupyterlab/security/advisories/GHSA-37w4-hwhx-4rc4) and [GHSA-89vp-jrxv-24w8](https://github.com/jupyterlab/jupyterlab/security/advisories/GHSA-89vp-jrxv-24w8), this does not need an allowlist or blocklist to be configured. The uninstall path never consulted the listing. ### Impact An authenticated user gains a read of server-side files and an outbound request from the server, both outside the confinement that the contents root and the disabled kernels were meant to provide. The user cannot choose which line of a file is returned, and cannot write content of their choosing. #### Reading files the account cannot reach `-r<path>` makes pip open the path as a requirements file. The first line that pip cannot parse as a requirement comes back in the response together with the path. Blank lines and `#` comments are skipped. A line that is a valid package name is accepted silently, and the request answers 201 with no message, so a secret made only of letters, digits, dots, hyphens and underscores is not disclosed. One line is returned per request, and the requester cannot select which one. The same option distinguishes a missing path (`[Errno 2] No such file or directory`), an unreadable one (`[Errno 13] Permission denied`) and a directory (`[Errno 21] Is a directory`), so any path on the host can be probed for existence and readability. #### Reaching hosts and ports the account cannot reach `-r` also accepts a URL. pip fetches it with an ordinary GET from the server's network position and reflects the first unparsable line of the response body the same way. A response whose body is a single line comes back in full. In a deployment where the single-user server sits inside a private network, this reaches internal services and cloud instance metadata endpoints that the user has no other route to. #### Writing files `--log=<path>` makes pip create that file if it is absent and append its own log to it otherwise. The content is pip's log text, not text the requester chooses, so the effect is to create files at chosen paths and to corrupt a file that has to parse, such as a configuration file read at the next start. #### What this does not add The same endpoint already uninstalls any installed package when it is given an ordinary, valid package name, so the injection adds no availability impact beyond what the extension manager already permits. Deployments that treat that as unacceptable should use the read-only extension manager, as described below. pip refuses `--python` after a subcommand name, and ignores editable entries in an uninstall requirements file, so neither gives code execution. ### Patches JupyterLab [`v4.6.4`](https://github.com/jupyterlab/jupyterlab/releases/tag/v4.6.4) and [`v4.5.11`](https://github.com/jupyterlab/jupyterlab/releases/tag/v4.5.11) contain the patch. The uninstall name is now validated as a PyPI package name in both the HTTP handler and `PyPIExtensionManager.uninstall`, and the pip command line carries an explicit `--` before the package operand. Users of applications that depend on JupyterLab, such as Notebook v7+, should update `jupyterlab` package too. ### Workarounds Deployments wanting to disable programmatic extension installation and removal entirely can switch to the read-only extension manager: ```bash --LabApp.extension_manager=readonly ``` or the following traitlet: ```python c.LabApp.extension_manager = 'readonly' ``` You can confirm that the read-only manager is in use from GUI: <img width="293" height="293" alt="image" src="https://github.com/user-attachments/assets/8016c809-633e-4ed0-a5bc-6bc4793caa0f" /> Users lose the ability to install and remove extensions from the Extension Manager; extensions must then be installed by an administrator in the environment. Note: an egress policy that blocks outbound requests from the single-user server limits the reach of the URL variant, but does not address the file read or the file write.

CVE-2026-102904 Affected JupyterLab Versions

2 component(s)
Fixed Versions:
4.5.114.6.4
Product / Package Ecosystem Affected Versions Fixed Version
jupyterlab
PyPI
≥ 4.0.0
< 4.5.11
4.5.11
jupyterlab
PyPI
≥ 4.6.0
< 4.6.4
4.6.4

References

Frequently Asked Questions

What is CVE-2026-102904?
CVE-2026-102904 is a security vulnerability in JupyterLab involving Argument injection in JupyterLab extension uninstall exposes server-readable files and internal URLs under CWE-88: Improper Neutralization of Argument Delimiters in a Command ('Argument Injection').
Which JupyterLab versions are affected?
Affected versions include releases prior to 4.5.11, 4.6.4 across affected packages: jupyterlab.
What version fixes CVE-2026-102904?
Upgrading to 4.5.11, 4.6.4 (or the latest patched release for your respective branch) resolves the vulnerability.
Is jupyterlab affected?
Yes, jupyterlab is affected in releases prior to 4.5.11, 4.6.4. Upgrade your dependency to 4.5.11, 4.6.4 or later.
What is the CVSS score of CVE-2026-102904?
CVE-2026-102904 has a CVSS v3.1 base score of 5.4 (MODERATE) with vector string CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:L/I:L/A:N.
What CWE is associated with CVE-2026-102904?
CVE-2026-102904 is classified under CWE-88: Improper Neutralization of Argument Delimiters in a Command ('Argument Injection') (The product constructs a string for a command to be executed by a separate component in another control sphere, but it does not properly delimit the intended arguments, options, or switches within that command string.).
How can I check PyPI dependencies for CVE-2026-102904?
Run the dependency tree command: 'pip show jupyterlab'. This command lists direct and transitive dependencies in your build configuration so you can verify if a vulnerable version is resolved.
Machine-readable OSV JSON OSV v1.5 Schema
Normalized schema output
{
  "schema_version" : "1.4.0",
  "id" : "GHSA-3325-v43h-43rv",
  "modified" : "2026-10-01T15:21:17Z",
  "published" : "2026-10-01T15:21:17Z",
  "aliases" : [ "CVE-2026-102904" ],
  "summary" : "JupyterLab: Argument injection in JupyterLab extension uninstall exposes server-readable files and internal URLs",
  "details" : "JupyterLab's PyPI extension manager runs `python -m pip uninstall` with the extension name taken from the request body. `ExtensionHandler.post` validates the name for `cmd=install` but not for `cmd=uninstall`, so a name that begins with `-` reaches the command line and pip reads it as an option rather than as a package.\n\n```python\ncmdline = [\n    sys.executable,\n    \"-m\",\n    \"pip\",\n    \"uninstall\",\n    \"--yes\",\n    \"--no-input\",\n    extension,\n]\n```\n\nAn extension name of the form `-r` followed by a path therefore made pip open that path as a requirements file. pip's parse error quotes the line it could not read and names the file it came from, and JupyterLab returned that error in the response body, so the line reached the requester.\n\nThis has security implications only for deployments that combine all of the following:\n\n- the (default) PyPI Extension Manager enabled, so that uninstall requests reach pip;\n- an authenticated account permitted to call the extension API; and\n- kernels and terminals disabled or delegated to remote hosts (otherwise a user with kernel access can read the same files and make the same outbound requests directly, regardless of this endpoint)\n\nUnlike [GHSA-37w4-hwhx-4rc4](https://github.com/jupyterlab/jupyterlab/security/advisories/GHSA-37w4-hwhx-4rc4) and [GHSA-89vp-jrxv-24w8](https://github.com/jupyterlab/jupyterlab/security/advisories/GHSA-89vp-jrxv-24w8), this does not need an allowlist or blocklist to be configured. The uninstall path never consulted the listing.\n\n### Impact\n\nAn authenticated user gains a read of server-side files and an outbound request from the server, both outside the confinement that the contents root and the disabled kernels were meant to provide. The user cannot choose which line of a file is returned, and cannot write content of their choosing.\n\n#### Reading files the account cannot reach\n\n`-r<path>` makes pip open the path as a requirements file. The first line that pip cannot parse as a requirement comes back in the response together with the path. Blank lines and `#` comments are skipped. A line that is a valid package name is accepted silently, and the request answers 201 with no message, so a secret made only of letters, digits, dots, hyphens and underscores is not disclosed. One line is returned per request, and the requester cannot select which one.\n\nThe same option distinguishes a missing path (`[Errno 2] No such file or directory`), an unreadable one (`[Errno 13] Permission denied`) and a directory (`[Errno 21] Is a directory`), so any path on the host can be probed for existence and readability.\n\n#### Reaching hosts and ports the account cannot reach\n\n`-r` also accepts a URL. pip fetches it with an ordinary GET from the server's network position and reflects the first unparsable line of the response body the same way. A response whose body is a single line comes back in full. In a deployment where the single-user server sits inside a private network, this reaches internal services and cloud instance metadata endpoints that the user has no other route to.\n\n#### Writing files\n\n`--log=<path>` makes pip create that file if it is absent and append its own log to it otherwise. The content is pip's log text, not text the requester chooses, so the effect is to create files at chosen paths and to corrupt a file that has to parse, such as a configuration file read at the next start.\n\n#### What this does not add\n\nThe same endpoint already uninstalls any installed package when it is given an ordinary, valid package name, so the injection adds no availability impact beyond what the extension manager already permits. Deployments that treat that as unacceptable should use the read-only extension manager, as described below.\n\npip refuses `--python` after a subcommand name, and ignores editable entries in an uninstall requirements file, so neither gives code execution.\n\n### Patches\n\nJupyterLab [`v4.6.4`](https://github.com/jupyterlab/jupyterlab/releases/tag/v4.6.4) and [`v4.5.11`](https://github.com/jupyterlab/jupyterlab/releases/tag/v4.5.11) contain the patch. The uninstall name is now validated as a PyPI package name in both the HTTP handler and `PyPIExtensionManager.uninstall`, and the pip command line carries an explicit `--` before the package operand.\n\nUsers of applications that depend on JupyterLab, such as Notebook v7+, should update `jupyterlab` package too.\n\n### Workarounds\n\nDeployments wanting to disable programmatic extension installation and removal entirely can switch to the read-only extension manager:\n\n```bash\n--LabApp.extension_manager=readonly\n```\n\nor the following traitlet:\n\n```python\nc.LabApp.extension_manager = 'readonly'\n```\n\nYou can confirm that the read-only manager is in use from GUI:\n\n<img width=\"293\" height=\"293\" alt=\"image\" src=\"https://github.com/user-attachments/assets/8016c809-633e-4ed0-a5bc-6bc4793caa0f\" />\n\nUsers lose the ability to install and remove extensions from the Extension Manager; extensions must then be installed by an administrator in the environment.\n\nNote: an egress policy that blocks outbound requests from the single-user server limits the reach of the URL variant, but does not address the file read or the file write.",
  "severity" : [ {
    "type" : "CVSS_V3",
    "score" : "CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:L/I:L/A:N"
  } ],
  "affected" : [ {
    "package" : {
      "ecosystem" : "PyPI",
      "name" : "jupyterlab"
    },
    "ranges" : [ {
      "type" : "ECOSYSTEM",
      "events" : [ {
        "introduced" : "4.6.0"
      }, {
        "fixed" : "4.6.4"
      } ]
    } ],
    "database_specific" : {
      "last_known_affected_version_range" : "<= 4.6.3"
    }
  }, {
    "package" : {
      "ecosystem" : "PyPI",
      "name" : "jupyterlab"
    },
    "ranges" : [ {
      "type" : "ECOSYSTEM",
      "events" : [ {
        "introduced" : "4.0.0"
      }, {
        "fixed" : "4.5.11"
      } ]
    } ],
    "database_specific" : {
      "last_known_affected_version_range" : "<= 4.5.10"
    }
  } ],
  "references" : [ {
    "type" : "WEB",
    "url" : "https://github.com/jupyterlab/jupyterlab/security/advisories/GHSA-3325-v43h-43rv"
  }, {
    "type" : "ADVISORY",
    "url" : "https://nvd.nist.gov/vuln/detail/CVE-2026-102904"
  }, {
    "type" : "WEB",
    "url" : "https://github.com/jupyterlab/jupyterlab/commit/9e1951e57da499ca66ef5e0caf68e71ad2e3d1a5"
  }, {
    "type" : "WEB",
    "url" : "https://github.com/jupyterlab/jupyterlab/commit/a274a8276b9185d03efd4c3c20713d3137ac49e6"
  }, {
    "type" : "WEB",
    "url" : "https://github.com/jupyterlab/jupyterlab/commit/e277bc958e737130ac47b6c5078d08b29298828f"
  }, {
    "type" : "PACKAGE",
    "url" : "https://github.com/jupyterlab/jupyterlab"
  }, {
    "type" : "WEB",
    "url" : "https://github.com/jupyterlab/jupyterlab/releases/tag/v4.5.11"
  }, {
    "type" : "WEB",
    "url" : "https://github.com/jupyterlab/jupyterlab/releases/tag/v4.6.4"
  } ],
  "database_specific" : {
    "cwe_ids" : [ "CWE-209", "CWE-88", "CWE-918" ],
    "severity" : "MODERATE",
    "github_reviewed" : true,
    "github_reviewed_at" : "2026-10-01T15:21:17Z",
    "nvd_published_at" : "2026-09-29T21:17:18Z"
  }
}