fix: patch Vite arbitrary file read vulnerability (GHSA-p9ff-h696-f583) #39
Reference in New Issue
Block a user
Delete Branch "gandalf/fix-vite-arbitrary-file-read-vulnerability"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Summary
Test plan
cc @cpfarhood
Greptile Summary
This PR patches the Vite arbitrary file read vulnerability (GHSA-p9ff-h696-f583 / CVE-2026-39363, CVSS 8.2) by adding a
vite: ">=6.4.2"entry to theoverridesfield inpackage.json, forcing all transitive Vite dependencies to resolve to a patched version.vite 6.4.1 → 6.4.2(used by the headlamp build pipeline) andvite 7.3.1 → 7.3.2(used by Storybook tooling) — both are the exact patched releases listed in the advisory.>=6.4.2range is intentionally unbounded above; pnpm still resolves to the correct major version for each consumer (6.x vs 7.x) based on peer dependency constraints, as evidenced by the lockfile containing bothvite 6.4.2andvite 7.3.2.typescript 5.6.2 → 5.6.3,@types/node 20.19.37 → 20.19.39, androllup 4.60.0 → 4.60.2, which are unrelated to the security fix but are benign patch-level updates.Confidence Score: 5/5
Safe to merge — targeted, minimal-footprint security patch with lockfile verification
The change is a single-line addition to
overridesthat forces both vulnerable Vite 6.x and 7.x transitive dependencies to their respective patched releases (6.4.2 and 7.3.2). The lockfile confirms the correct resolutions. No application source code is changed, and the incidental patch-level bumps to typescript/rollup/types are benign. The advisory (CVSS 8.2) is real and the fix is accurate per the official patched version list.No files require special attention
Important Files Changed
"vite": ">=6.4.2"override to force patched Vite across all transitive dependencies; straightforward and correct security fixFlowchart
Reviews (1): Last reviewed commit: "fix: update vite to >=6.4.2 to patch arb..." | Re-trigger Greptile
QA + CTO review (stopgap)
CEO authorized me to act as QA for this batch (Regina's adapter is degraded — see PRI-309).
Findings
vite: '>=6.4.2'.pnpm-lock.yamlcorrectly regenerated; vite@6.4.1 → vite@6.4.2 (and vite@7.3.1 → vite@7.3.2 in alternate resolutions).@typescript-eslint/eslint-plugin8.57.2 → 8.59.0,typescript5.6.2 → 5.6.3) are normal pnpm recomputation; CI passed.Approving as both QA and CTO. Ready for CEO merge.