agentleFS
rpraharajrpraharaj/ai-engineer/skills/upgrade-deps/SKILL.md

upgrade-deps

Skill1 starsChanged 16 days agoMITView on GitHub

Update dependencies with this repo's package manager, run the test suite, and fix breakages in this repo's code. Use when the user asks to upgrade dependencies, bump packages, apply dependabot-style updates, or refresh the lockfile. Do not hide breaking changes.

11 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use create-pr.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • SentrySkill

    create-pr

    Create pull requests following Sentry conventions. Use when opening PRs, writing PR descriptions, or preparing changes for review. Follows Sentry's code review guidelines.

    Also in 3 projects that use this

The file

Show its rules and text

We split this skill into the rules it gives an agent

15 rules. 0 are common, 9 are distinctive, and 6 are very specific to this project.

  • 0

    Common

    Most of the 11 edited versions we read keep these.

  • 9

    Distinctive

    This author's own choices. Worth a look.

  • 6

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  • could be automatic: a setting could enforce it
  1. Distinctive:

    Read AGENTS.md if present.

    line 14
  2. Distinctive:

    Use the lockfile / manifest that exists.

    line 18
  3. Distinctive:

    If several ecosystems exist (monorepo), stay in the package the user named, or one ecosystem per run.

    line 40
  4. Distinctive:

    Default: the packages the user named.

    line 44
  5. Very specific:

    Prefer this repo's existing upgrade path (Renovate/Dependabot config, a make deps target, npm outdated)

    line 46
  6. Distinctive:

    Write the planned list before editing: name, from → to, major bump yes/no.

    line 49
  7. Distinctive:

    Do not upgrade a package that is vendored or pinned in AGENTS.md Do not edit without asking.

    line 51
  8. Distinctive:

    Run the manager's update command for the planned set.

    line 56
  9. Distinctive:

    Do not hand-edit a lockfile if the manager can write it.

    line 57
  10. Very specific:

    Do not add new packages here (implement-change does that).

    line 58
  11. Very specific:

    Run Test, then Lint / Typecheck / Build if known.

    line 62
  12. Very specific:

    Follow debug if the cause is unclear

    line 67
  13. Distinctive:

    Do not delete tests, skip cases, or pin the old version in a new override without telling the user

    line 68
  14. Very specific:

    If the upgrade is a major and the fix is large, stop, revert the dep bump, and report the breakage.

    line 69
  15. Very specific:

    Do not commit unless asked.

    line 85

“Common” is measured: kept by at least 6 in 10 of the 11 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

Upgrade deps

Move dependencies forward without lying about compatibility. Do not swallow a breaking change by weakening tests or pinning * ranges this repo does not already use.

Read AGENTS.md if present.

Detect the manager

Use the lockfile / manifest that exists. Do not add a second manager.

EvidenceManager
pnpm-lock.yamlpnpm
yarn.lockyarn
bun.lock / bun.lockbbun
package-lock.jsonnpm
uv.lock / poetry.lock / Pipfile.lockuv / poetry / pipenv
requirements.txt + no other lockpip (say it is weaker)
Cargo.lockcargo
go.sumgo
Gemfile.lockbundler
composer.lockcomposer
pom.xml / maven-wrapper.propertiesmaven
build.gradle / build.gradle.kts / gradle.lockfilegradle
*.csproj / packages.lock.json / Directory.Packages.propsdotnet / nuget
Package.resolvedswiftpm
pubspec.lockdart / flutter (pub)
flake.locknix
.terraform.lock.hclterraform

If several ecosystems exist (monorepo), stay in the package the user named, or one ecosystem per run.

Scope

Default: the packages the user named. If they said "all" or "latest":

  • Prefer this repo's existing upgrade path (Renovate/Dependabot config, a make deps target, npm outdated)
  • Otherwise update within the same major unless the user asked for majors

Write the planned list before editing: name, from → to, major bump yes/no.

Do not upgrade a package that is vendored or pinned in AGENTS.md Do not edit without asking.

Apply

  1. Record the current lockfile hash / git status (clean tree preferred; if dirty, only touch dep files or stop).
  2. Run the manager's update command for the planned set. Match commands this repo already uses in CI or README.
  3. Do not hand-edit a lockfile if the manager can write it.
  4. Do not add new packages here (implement-change does that). This skill only changes versions of existing deps.

Prove

Run Test, then Lint / Typecheck / Build if known. If a command is (unknown), say unverified for that check.

On failure:

  • Fix this repo's call sites (types, renamed APIs, import paths)
  • Follow debug if the cause is unclear
  • Do not delete tests, skip cases, or pin the old version in a new override without telling the user
  • If the upgrade is a major and the fix is large, stop, revert the dep bump, and report the breakage. Offer shape-work for an adoption change

Load engineering-bar (paths in that skill) on any product-code edits.

Done

## Upgrade
- Manager:
- Changes: <name from → to>
- Majors:
- Commands: <test/lint + exit status>
- Product-code fixes:
- Left on old version (and why):

Do not commit unless asked. Next: ready-to-merge if product code changed, then commit. Mention ship-notes if this repo logs dependency bumps in the changelog.

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.