agentleFS
spleenteospleenteo/maestro/.claude/skills/maintenance/SKILL.md

maintenance

Skill1 starsChanged 46 days agoMITView on GitHub

Routine maintenance of the Maestro template repository, run from the repository itself and never from an instance. Runs the read-only report `.claude/skills/maintenance/check` (Python suite, mod tests, plugin manifests, GitHub Actions pins, Pages run, Dependabot PRs, merged branches, brew tools, sqlite-vec pin, installed plugin cache), explains each finding in plain words and applies the standard remedy from `docs/maintenance.md` on the owner's yes. Use when the owner types /maintenance, says "manutenzione", "facciamo manutenzione", "è tutto aggiornato?", "maintenance", "is everything up to date", or after an update of Claude Code.

  • Commits and pushes

18 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use recall and release.

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.
  • termdockSkill

    recall

    Retrieves cross-session memories about past decisions, lessons, and patterns. Use when user asks about prior work, past decisions, or mentions a feature. Triggers include check memory, what did we decide, how did we solve.

    Also in 3 projects that use this

  • MicrosoftSkill

    release

    Use ONLY when the user explicitly asks to cut/publish a Huabu desktop release or tag a version. Performs an outward action — bumps the version in apps/desktop/package.json, commits + annotated-tags vX.Y.Z, and pushes the branch + tag. That tag push auto-builds the macOS/Windows installers and publishes a GitHub Release. Provides `status`, `bump`, `tag-push`, and a one-command `run` (bump → tag-push).

    Also in 3 projects that use this

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 14 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:

    Only in the template repository: plugins/maestro/ present at the root.

    line 13
  2. Very specific:

    After the remedies, rerun check --no-tests and show the summary line.

    line 50

“Common” is measured: kept by at least 6 in 10 of the 14 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

Maintenance

The template repository has no lockfile. Its dependencies are Claude Code (hook events, skill frontmatter, claude plugin commands, the mod API), the GitHub Actions of pages.yml, sqlite-vec through uv, and the external tools (yap, ollama, sqlite3, rg). This skill checks all of them in one pass and keeps the remedies in docs/maintenance.md, so the procedure is never rebuilt from scratch.

Where it runs

Only in the template repository: plugins/maestro/ present at the root. In an instance or a satellite, say in one sentence that maintenance belongs to the template repository and stop.

1. Run the report

.claude/skills/maintenance/check              # full: the Python suite takes about four minutes
.claude/skills/maintenance/check --no-tests   # when the owner asks for a quick look

Without instructions, run the full report. When the owner asks for a quick look ("veloce", "quick", "solo un'occhiata"), pass --no-tests and tell them in the report that the test suites were skipped.

The full report takes about five minutes and the Bash tool stops a command after two: set the tool's timeout parameter to 600000 ms or more, and launch the script as it is, in a single call.

The exit code is a signal, never an error of the script: 1 means the report has at least one ATTENTION line to bring to the owner, 0 means every check is OK or SKIP.

2. Read the remedies

Read docs/maintenance.md: the table says what each check means, the standard remedy, and which decisions the owner has deferred on purpose (those are reported as open, never "fixed"). Check the auto-memory for a maintenance entry of the last month, so a decision taken then is not asked again.

3. Report to the owner

In chat, in the owner's language, grouped in three blocks: healthy, needs attention, skipped (with the reason). Plain words, no internal jargon: "the deploy workflow uses versions GitHub is retiring" rather than the raw line. Figures stay (test counts, versions, branch names). For each ATTENTION line, state the remedy from docs/maintenance.md and whether it is a repo change, a machine change (the owner's brew), or a decision.

4. Apply on yes

Each remedy waits for the owner's yes; one yes covers the remedies named in the same question. Rules:

  • Repo changes follow docs/development-guidelines.md: explicit git add on the touched files, one commit per kind (chore(ci): …, chore(repo): …), English imperative message. Commit on a branch when on main, unless the owner says otherwise. Never push, merge or delete remote branches without an explicit request.
  • Actions bump: edit the @vN pins in .github/workflows/*.yml to the latest major the report names; the workflow verifies itself at the next push to main that touches site/, CHANGELOG.md or the workflow.
  • Dependabot PR: read its diff with gh pr diff <n> before proposing the merge; merge only on yes.
  • Merged branches: git branch -d for local, git push origin --delete for remote, only the ones the report lists.
  • Brew tools: never run brew upgrade; give the command to the owner.
  • Plugin behind main: claude plugin update <name>@maestro runs on yes; stale cached versions need nothing.
  • version in plugin.json: a deferred decision, see docs/maintenance.md; never add it inside this skill.

5. Close

After the remedies, rerun check --no-tests and show the summary line. When the test count changed, update the baseline line in docs/maintenance.md and in docs/development-guidelines.md → ## Gate. Write one auto-memory with the date, what was found and what was done.

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.