dependency-gardening
Skill0 starsChanged 24 days agoMITView on GitHub
Refresh dependency versions and their adoption across the repo. Use when upgrading packages, reviewing dependency update PRs, or reassessing how we use a library.
In this repository2 other files it gives its agents.
- feedback-loopsSkill
- keyword-first-seoSkill
The file
We split this skill into the rules it gives an agent
4 rules. About 0 are common, 4 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈4
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 8
Keep dependencies current and their adoption idiomatic, simple, and efficient.
- Distinctive: line 16
Read upstream: Read current official documentation and recommended practices, plus the full changelog and migration guides between our …
- Distinctive: line 27
Verify and report: Follow repository dependency conventions and validation gates.
- Distinctive: line 31
Every usage is accounted for: improved, retained with a reason, or explicitly deferred.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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.
Files
Dependency Gardening
Keep dependencies current and their adoption idiomatic, simple, and efficient. Each pass reassesses our entire use of the selected library with today's knowledge and capabilities. Release notes are one input; existing code deserves fresh scrutiny even when upstream hasn't changed.
Work on the selected dependency or coupled family. This is intentional maintenance, not a schedule to update everything every day.
Workflow
- Establish the current state. Identify installed versions, all consumers, and the latest stable release. Prefer staying current; explain concrete constraints when choosing an older target. Treat tightly coupled packages as one upgrade.
- Read upstream. Read current official documentation and recommended practices, plus the full changelog and migration guides between our installed and target versions. Include intermediate releases. Follow linked details that could affect our usage; state documentation gaps.
- Reassess the whole adoption. Inspect every usage across the repo: callers, configuration, wrappers, abstractions, patches, and workarounds. Evaluate the approach afresh rather than limiting changes to the release delta. Look for:
- Misused APIs, outdated patterns, and missed built-in capabilities.
- Unnecessary wrappers, duplicated logic, and avoidable complexity.
- Correctness, lifecycle, and performance problems.
- Workarounds and overrides whose justification no longer holds.
- Inconsistent patterns across consumers.
- Choose and implement improvements. Recommend adopt now, retain, or defer, grounded in concrete code and upstream evidence. Prefer the simplest sound implementation for our actual needs. Include clear improvements within the approved scope; route consequential redesigns through the existing planning process. Verify that upstream behavior covers our requirements before removing workarounds. Keep good existing code; novelty alone earns no rewrite.
- Verify and report. Follow repository dependency conventions and validation gates. Exercise affected behavior and measure performance when claiming a local improvement. Report versions, source links, improvements adopted, code removed, worthwhile deferrals, and checks run. Distinguish upstream claims from locally verified benefits.
Completion
Every usage is accounted for: improved, retained with a reason, or explicitly deferred. Group equivalent usages in the report. A version-only update—or no change—is valid when the full assessment supports it.
Sources
- Florian's dependency-gardening brief, 2026-10-20: stay current and reassess the whole adoption with each pass, beyond changelog-driven fixes.
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
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.