Skip to content

deps: bump gitpython 3.1.50 -> 3.1.52 in the production lockfile (clears 3 GHSA advisories)#151

Merged
Gavin-Borges merged 2 commits into
mainfrom
deps/lock-gitpython-3.1.52
Jul 24, 2026
Merged

deps: bump gitpython 3.1.50 -> 3.1.52 in the production lockfile (clears 3 GHSA advisories)#151
Gavin-Borges merged 2 commits into
mainfrom
deps/lock-gitpython-3.1.52

Conversation

@Gavin-Borges

Copy link
Copy Markdown
Owner

What

Bumps gitpython 3.1.50 -> 3.1.52 in environments/requirements.lock (transitive via snakemake).

Why

The production lockfile still pinned the vulnerable gitpython==3.1.50 even though environments/requirements-ci.txt was already bumped to 3.1.52 (#136 only touched the CI manifest; the production lock lagged). 3.1.50 carries three advisories:

  • GHSA-2f96-g7mh-g2hx - command injection via long-option prefix abbreviation
  • GHSA-956x-8gvw-wg5v
  • GHSA-v396-v7q4-x2qj

All are resolved in 3.1.51+. Actual runtime risk is low (gitpython is not on the peptide-scoring path), but a fix exists, so per project policy we bump rather than dismiss.

Verification

  • Wheel + sdist sha256 hashes match PyPI for 3.1.52 and the existing requirements-ci.txt pins exactly (same version, same two hashes).
  • No manifest still references 3.1.50 (grep gitpython==3.1.5 -> only 3.1.52).
  • 3-stage pre-push gate green (selftest 18/18; integrity 20P/1W/0F/2SKIP; fast tests pass).

Note on end-to-end proof

This was surfaced by auditing the installed lockfile with pip-audit (PR #148). Because this branch is off main (which does not yet carry that pip-audit fix), this branch's advisory scan is still the old lock-resolve form. The definitive "0 gitpython findings" proof lands once both this PR and #148 are on main; recommended order is to merge #148 first, then update this branch.

environments/requirements.lock still pinned gitpython 3.1.50 (transitive via
snakemake) while environments/requirements-ci.txt was already at 3.1.52, so the
production lock lagged the fix. 3.1.50 carries GHSA-2f96-g7mh-g2hx,
GHSA-956x-8gvw-wg5v and GHSA-v396-v7q4-x2qj (command injection via long-option
prefix abbreviation), all resolved in 3.1.51 and later. The wheel and sdist
hashes match PyPI and the existing ci manifest. Found by auditing the installed
lockfile with pip-audit.
@Gavin-Borges
Gavin-Borges merged commit d112527 into main Jul 24, 2026
17 of 18 checks passed
@Gavin-Borges
Gavin-Borges deleted the deps/lock-gitpython-3.1.52 branch July 24, 2026 03:22
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.

1 participant