agentleFS
bekalpaslanbekalpaslan/promptline/.claude/skills/maintenance/SKILL.md

maintenance

Skill13 starsChanged 6 days agoMITView on GitHub

Run a maintenance pass on Promptline, keeping what exists healthy and true with no features, redesigns or behaviour changes. Use this whenever the user asks for maintenance, upkeep, housekeeping, chores, a tidy-up or cleanup of the repo, or asks to get the project back in shape between releases, even without the word maintenance. It checks the baseline, then runs the dependency and doc skills, prunes the backlog, refreshes the architecture map if it's due, and lists what's left to clean up, all on a chore branch, and reports what changed and what needs the maintainer. For an open-ended review of what to improve, use fresh-eyes instead. It never releases.

18 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use memory and recall.

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.
  • GoogleSkill

    memory

    Records and appends every user-agent customer support interaction to the memory.md log in the workspace.

    Also in 3 projects that use this

  • 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

The file

We split this skill into the rules it gives an agent

11 rules. 0 are common, 7 are distinctive, and 4 are very specific to this project.

  • 0

    Common

    Most of the 14 edited versions we read keep these.

  • 7

    Distinctive

    This author's own choices. Worth a look.

  • 4

    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:

    Check the git state first, and never disturb work in progress: if the checkout is dirty or on a branch with open work, use a fresh worktree …

    line 21
  2. Very specific:

    Dependencies: Run /dependabot and follow its standing rules.

    line 31
  3. Very specific:

    Docs: Run /doc-audit, scoped to everything since the latest release tag.

    line 32
  4. Distinctive:

    Ask before deleting anything.

    line 41
  5. Distinctive:

    Follow the conventions, line-ending rules and lockfile recipes in CLAUDE.md.

    line 45
  6. Very specific:

    If a fix must change behaviour, update BEHAVIOR.md in the same commit and add a test named for the behaviour.

    line 47
  7. Distinctive:

    Leave generated media alone unless the pass changed what it shows.

    line 51
  8. Distinctive:

    Don't release, tag, bump the version, touch the updater feed, or post anywhere.

    line 55
  9. Distinctive:

    Don't rewrite history or change signing keys or identifiers.

    line 57
  10. Very specific:

    Don't merge to master or push without the maintainer's go-ahead.

    line 58
  11. Distinctive:

    Run all the checks again from the branch and remove any worktrees you created.

    line 65

“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

A pass that keeps what exists healthy and true: no features, no redesigns, no behaviour changes. CLAUDE.md is the authority on how this repo works, and this skill follows it rather than restating it. Work out the current state yourself, and start from that and BACKLOG.md, not from a new audit.

Scope

$ARGUMENTS narrows the pass to one step of the list below. Empty means all of them, in order, skipping any that is already clean.

Setup

Check the git state first, and never disturb work in progress: if the checkout is dirty or on a branch with open work, use a fresh worktree off an up-to-date master (CLAUDE.md → Working in parallel has what a new worktree needs). Do the work on a chore/maintenance-<date> branch.

The pass

  1. Baseline. Confirm CI is green on master and run every check in CLAUDE.md → Checks. Fix what's red at the cause, without suppressing, skipping or loosening anything.
  2. Dependencies. Run /dependabot and follow its standing rules.
  3. Docs. Run /doc-audit, scoped to everything since the latest release tag.
  4. Backlog. Flag entries that are already done or overtaken by a decision. Do only small, safe entries that don't change behaviour.
  5. Architecture map. Refresh it if the rule in CLAUDE.md → Architecture map applies, in its own commit.
  6. Housekeeping. List finished worktrees, merged branches, and anything with a date coming due, such as the certificates and keys under the custody notes. Keep branches that hold shelved work, theme/* above all. Ask before deleting anything.

Standards

  • Follow the conventions, line-ending rules and lockfile recipes in CLAUDE.md. Keep one concern per commit.
  • If a fix must change behaviour, update BEHAVIOR.md in the same commit and add a test named for the behaviour.
  • Anything out of scope goes in BACKLOG.md rather than widening the change.
  • Leave generated media alone unless the pass changed what it shows.

Limits

  • Don't release, tag, bump the version, touch the updater feed, or post anywhere.
  • Don't rewrite history or change signing keys or identifiers.
  • Don't merge to master or push without the maintainer's go-ahead. That holds inside /dependabot too.
  • Bring decisions that are the maintainer's one at a time, each with a recommended default.

Finish

Run all the checks again from the branch and remove any worktrees you created. Report what changed by commit, what you skipped and why, what you added to the backlog, and what needs the maintainer's decision. Show failing output in full.

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.