Skip to content

Clear the open dependency vulnerabilities - #11

Merged
matt-humaan merged 1 commit into
masterfrom
update-clear-soon-vulnerabilities-sup-3894
Aug 14, 2026
Merged

Clear the open dependency vulnerabilities#11
matt-humaan merged 1 commit into
masterfrom
update-clear-soon-vulnerabilities-sup-3894

Conversation

@matt-humaan

Copy link
Copy Markdown
Contributor

One of five projects Vanta monitors, under SUP-3894. This repo has two lockfiles and both were flagged. Everything here is transitive and dev-scope — nothing reaches consumers of the package — but it is what Vanta is flagging.

Root

Package Change Advisory
brace-expansion 5.0.7 → 5.0.9 GHSA-rgw5-rvv9-x895, GHSA-mh99-v99m-4gvg
js-yaml 3.15.0 → 3.15.1 GHSA-5p4m-2wfm-xmqj
tar 7.5.20 → 7.5.22 GHSA-r292-9mhp-454m
undici 7.28.0 → 7.29.0 GHSA-4cwx-7wf7-3272, plus four medium

demo/

nanoid 3.3.16 → 3.3.18 and postcss 8.5.20 → 8.5.26.

demo/package-lock.json is easy to miss — Dependabot reports it as a separate manifest, and a root-only sweep says the repo is clean while it isn't.

A lockfile refresh reaches all six, so there is no manifest change and no overrides block.

One apparent finding that isn't real

cbor2@2.3.0, pulled in by ava, looked like a high. It is a false positive: that advisory covers the PyPI package, and its range (<= 5.8.0) sits above every version cbor2 has ever published on npm, where the latest is 2.3.0. It surfaced because my first sweep matched advisories by package name without constraining the ecosystem. No change made — and npm rejecting the attempted override with notarget is what prompted the second look.

Verification

Every package in both lockfiles (353 and 68) resolved against OSV, rather than reading the audit summary. Both clean afterwards. Full suite passes: 95 tests, 99% statement coverage.

Worth recording why it was done that way: my first pass used per-package GitHub advisory API calls, which silently hit the 5,000/hr rate limit and returned nothing — which the script read as "clean". That is a false negative in the dangerous direction. The batched OSV query has no such failure mode and aborts rather than reporting clean.

🤖 Generated with Claude Code using Claude Opus 5 (1M context)

This repo has two lockfiles and both were flagged. Everything here is
transitive and dev-scope — nothing reaches consumers of the package — but
it is what Vanta is flagging.

Root:

  brace-expansion  5.0.7  -> 5.0.9   GHSA-rgw5-rvv9-x895, GHSA-mh99-v99m-4gvg
  js-yaml          3.15.0 -> 3.15.1  GHSA-5p4m-2wfm-xmqj
  tar              7.5.20 -> 7.5.22  GHSA-r292-9mhp-454m
  undici           7.28.0 -> 7.29.0  GHSA-4cwx-7wf7-3272 and four medium

demo/:

  nanoid           3.3.16 -> 3.3.18
  postcss          8.5.20 -> 8.5.26

A lockfile refresh reaches all of them, so no manifest change and no
overrides block.

A fifth apparent root finding, cbor2 via ava, was a false positive: that
advisory covers the PyPI package, and its range (<= 5.8.0) is above every
version cbor2 has ever published on npm, where the latest is 2.3.0. No
change made.

Verified by resolving every package in both lockfiles against the GitHub
advisory API, constrained to the npm ecosystem — that constraint is what
surfaced the cbor2 mismatch — rather than trusting the audit summary.
Both come back clean.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
@linear-code

linear-code Bot commented Aug 14, 2026

Copy link
Copy Markdown

SUP-3894

@matt-humaan
matt-humaan merged commit ae35641 into master Aug 14, 2026
6 checks passed
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