agentleFS
nuncaeslupusnuncaeslupus/literary-critic/.claude/skills/dep-upgrade/SKILL.md

dep-upgrade

Skill0 starsChanged 6 days agoMITView on GitHub

Upgrades a uv project's dependencies behind pip-audit and tests. Use when the user wants dependencies upgraded or vuln-checked. Not for adding one (uv add).

4 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

2 rules. 0 are common, 0 are distinctive, and 2 are very specific to this project.

  • 0

    Common

    Most of the 4 edited versions we read keep these.

  • 0

    Distinctive

    This author's own choices. Worth a look.

  • 2

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Prefer --upgrade-package when chasing one CVE or feature, because a repo-wide upgrade folds dozens of changes into one diff and hides the …

    line 31
  2. Very specific:

    Run it through uv run so it audits the project's environment.

    line 52

“Common” is measured: kept by at least 6 in 10 of the 4 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

dep-upgrade

Upgrades a uv-managed project's dependencies under a safety net: snapshot, lock, audit for CVEs, gate on the test suite, then classify what broke.

CANARY: dep-upgrade-loaded-2026-06-04-ea39c2b5-719c61e6e84e1453

When to load

Load for a deliberate upgrade of a uv project — refreshing the lockfile, chasing CVEs, pulling newer versions. Adding one package is a plain uv add; tooling scaffolding belongs to python-bootstrap.

Step 1 — Snapshot, then upgrade

cp uv.lock uv.lock.bak   # keeps the change reviewable and revertible
uv lock --upgrade        # or --upgrade-package NAME for one targeted bump
uv sync                  # lock rewrites the file only; tests need the new install

Prefer --upgrade-package when chasing one CVE or feature, because a repo-wide upgrade folds dozens of changes into one diff and hides the culprit.

Step 2 — See what changed

python3 "${CLAUDE_SKILL_DIR}/scripts/compare_lockfile.py" uv.lock.bak uv.lock

It splits changed into direct (declared in pyproject.toml) and transitive, with counts, plus added and removed — read this rather than the raw git diff uv.lock, which is mostly hash noise. Read the transitive bumps: they can pass locally on a cached wheel and fail a clean install, so re-resolve in a clean environment before trusting green.

Step 3 — Audit for CVEs

uv run --with pip-audit pip-audit

Run it through uv run so it audits the project's environment; a bare uvx pip-audit audits only its own isolated tool environment. Answer a finding by moving forward to the patched release; pinning back onto the vulnerable version reintroduces it. Surface any advisory with no fixed version as a risk.

Step 4 — Test gate and classify

make test        # or: uv run pytest

Green: report the upgrade with the direct/transitive breakdown and any CVE findings. Red: classify each failure as a behaviour change in a direct dep (read its changelog), a transitive bump exposing a latent bug, or a real incompatibility. Pin the minimum necessary in pyproject.toml, re-lock and re-run; fix the code or the pin rather than the test.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.