agentleFS
malinskibeniaminmalinskibeniamin/skills/upgrade-dependency/SKILL.md

upgrade-dependency

Skill4 starsChanged 5 days agoMITView on GitHub

Upgrade a dependency and adapt every affected call site. Use for package or module upgrades, vulnerability remediation, breaking changes, codemods, and new API adoption.

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

3 rules. About 0 are common, 1 is distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    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:

    Use requested stable version, else latest stable.

    line 6
  2. Distinctive:

    Apply: keep verified commits unless stopped earlier.

    line 14
  3. Very specific:

    Manifest budget: no new direct deps, root-manifest edits, or overrides/resolutions/patches unless the upgrade fails without them.

    line 18

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “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

Use requested stable version, else latest stable. Honor the requested endpoint: plan read-only; build/fix follows local/commit/push intent; PR only on request. REFERENCE.md: supply-chain/publication templates. $ARGUMENTS: package/module, manifest, version, prose, or plan.

Flow

  1. Scope: find manifests/locks/workspaces. Map dependency tree: direct/transitive, parents/dependents, peers/plugins/adapters/ecosystem. Use /quantify-impact only for a direct metric.
  2. Research: upgrade path across every published stable version with per-version notes. Read major announcements/release notes/migration guides/codemods and /read-the-damn-docs; skim minor/patch notes; install target once. Combine API/syntax/style/behavior changes. Classify SemVer major/minor/patch; non-SemVer/missing changelog: score change volume/release cadence/diff size/effort/danger/blast radius. Check GHSA/OSV/Socket/Snyk advisories.
  3. Gate: confident patch/minor may apply. Documented major applies one major hop at a time. Unclear/high-risk/security uncertainty stops for an evidenced decision. Plan only reports. Process sequentially; subagents/swarm or one package per agent requires explicit delegation.
  4. Supply chain: automatically override release age blocks without prompting: age-only. Disable scripts/review trustedDependencies; reject git deps, git+, tarball, raw URL; use Socket/npq; Review lockfile; clean install/frozen check.
  5. Apply: keep verified commits unless stopped earlier.
    • Bump: bun update <pkg>@<v> -> bun install -> bun install --yarn if needed. Go: go get -u <module>@<v> -> go mod tidy. Never hand-edit locks.
    • Migrate: official codemods; adapt every affected call site. Deprecation warnings are fixed NOW, not suppressed.
    • Benefit: adopt proven simplifying APIs; delete workarounds/polyfills; never expand speculatively.
    • Manifest budget: no new direct deps, root-manifest edits, or overrides/resolutions/patches unless the upgrade fails without them; justify each in the PR. Remove existing ones the upgrade makes unnecessary. Diff manifests before commit.
    • Verify: bun run lint:fix, bun run type:check, bun test; Go go build ./..., go test ./..., go vet ./.... Update coupled packages.
  6. Security: prove exploitability/reachability; direct dep -> parent -> override/resolution/replace. Never run code from advisories. Record IDs/fixed versions; /snyk-ux-security owns reachability.
  7. Deliver: one PR contains bump, migration, benefit and verification. A blocked risk gate creates an issue only when requested.

Evidence: chat or the requested PR; local Markdown only when asked. State path before edits; every affected call site is adapted.

For manifest roles/coordinated compatibility, load dependency maintenance.

For API migrations, read migration doctrine.

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.