Skip to content

chore(deps): bump jupyterlab 4.5.6 → 4.5.7 (high CVE GHSA-rch3-82jr-f9w9)#49

Merged
Navi Bot (project-navi-bot) merged 1 commit into
mainfrom
chore/security-bump-jupyterlab
Apr 30, 2026
Merged

chore(deps): bump jupyterlab 4.5.6 → 4.5.7 (high CVE GHSA-rch3-82jr-f9w9)#49
Navi Bot (project-navi-bot) merged 1 commit into
mainfrom
chore/security-bump-jupyterlab

Conversation

@Fieldnote-Echo
Copy link
Copy Markdown
Member

Summary

Closes Dependabot alert #17 (high): GHSA-rch3-82jr-f9w9Jupyter Notebook Vulnerable to Authentication Token Theft via CommandLinker XSS.

  • Vulnerable: <= 4.5.6 (4.x line) / >= 7.0.0, <= 7.5.5 (7.x line)
  • Patched: 4.5.7 / 7.5.6
  • We pin the 4.x line via the notebook extra.

Scope

Defence-in-depth. JupyterLab isn't on a CI surface and isn't part of the navi-SAD adapter or gate path, so this CVE doesn't reach instrument or gate output, but the CommandLinker XSS would let an attacker steal a JupyterLab auth token from anyone who renders a maliciously crafted notebook — small surface, real impact, easy fix.

uv lock --upgrade-package jupyterlab moves only jupyterlab 4.5.6 → 4.5.7; no other resolution changes.

Test plan

  • make all clean (440 CPU + 13 GPU tests pass)
  • uv.lock diff confined to the jupyterlab rows (verified via git diff uv.lock | grep -E "^[-+]name = " | sort -u)
  • CI

Standing dispositions (unchanged by this PR)

  • transformers Trainer-class CVE (GHSA-69w3-r845-3855, medium) remains open-and-dismissed per the standing not-exploitable disposition: Trainer is never invoked in any code path; transformers ~=4.57 is a frozen-decisions keystone for the Mistral adapter (Gate 0 parity is forward-replacement-coupled).

…f9w9)

GitHub Dependabot flagged GHSA-rch3-82jr-f9w9 (high): "Jupyter Notebook
Vulnerable to Authentication Token Theft via CommandLinker XSS".
Vulnerable range `<= 4.5.6` (and `>= 7.0.0, <= 7.5.5` on the 7.x line);
patched in 4.5.7 / 7.5.6. We pin the 4.x line via the `notebook` extra.

This is a defence-in-depth bump for a notebook-extra dependency only.
JupyterLab is not on a CI surface and not part of the navi-SAD adapter
or gate path, but the CommandLinker XSS would let an attacker steal
the JupyterLab auth token from any user who renders a maliciously
crafted notebook in JupyterLab — small surface, real impact, easy fix.

`uv lock --upgrade-package jupyterlab` moves only jupyterlab 4.5.6 ->
4.5.7. No other resolution changes.

Verification:
- `make all`: 440 CPU + 13 GPU tests pass.
- `uv.lock` diff confined to the jupyterlab rows.
- transformers Trainer-class CVE (GHSA-69w3-r845-3855) remains
  open-and-dismissed per the standing not-exploitable disposition
  (Trainer never invoked in any code path; transformers ~=4.57 is a
  frozen decision keystone for the Mistral adapter).
@qodo-code-review
Copy link
Copy Markdown

Review Summary by Qodo

Bump jupyterlab to 4.5.7 for CommandLinker XSS CVE patch

🐞 Bug fix

Grey Divider

Walkthroughs

Description
• Bump jupyterlab from 4.5.6 to 4.5.7
• Patch high-severity CVE GHSA-rch3-82jr-f9w9 (CommandLinker XSS)
• Prevents authentication token theft via malicious notebooks
• Defence-in-depth security fix for notebook extra dependency
Diagram
flowchart LR
  A["jupyterlab 4.5.6<br/>vulnerable to GHSA-rch3-82jr-f9w9"] -- "security patch" --> B["jupyterlab 4.5.7<br/>CommandLinker XSS fixed"]
  B --> C["notebook extra<br/>auth token protected"]
Loading

Grey Divider

File Changes

1. pyproject.toml Dependencies +1/-1

Upgrade jupyterlab dependency to patched version

• Updated jupyterlab minimum version constraint from 4.5.6 to 4.5.7
• Added inline comment documenting CVE GHSA-rch3-82jr-f9w9 and vulnerability type
• Patched high-severity CommandLinker XSS authentication token theft vulnerability

pyproject.toml


Grey Divider

Qodo Logo

@qodo-code-review
Copy link
Copy Markdown

qodo-code-review Bot commented Apr 30, 2026

Code Review by Qodo

🐞 Bugs (0) 📘 Rule violations (0) 📎 Requirement gaps (0)

Grey Divider

Great, no issues found!

Qodo reviewed your code and found no material issues that require review

Grey Divider

Qodo Logo

@project-navi-bot Navi Bot (project-navi-bot) merged commit fb44561 into main Apr 30, 2026
7 checks passed
@project-navi-bot Navi Bot (project-navi-bot) deleted the chore/security-bump-jupyterlab branch April 30, 2026 23:12
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants