agentleFS
quietmachineworksquietmachineworks/qmw/skills/full-cycle/SKILL.md

full-cycle

Skill0 starsChanged 2 months agoMITView on GitHub

The full review cycle in one guided passage - audit-codebase, then a refactor per finding you pick, upgrade-deps, run-qa, and a check-release that ends on go or no-go for the tag

2 other projects use a version of this.

Adopted by

Most recently taken up by

  • Other versions2 projects use a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 2

    1. anotherben/claude-harness★ 7same name
    2. Mark-Life/agent-skills★ 4same name

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

  • In this repository12 other files it gives its agents.

The file

We split this skill into the rules it gives an agent

5 rules. About 0 are common, 1 is distinctive, and 4 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.

  • 4

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Read .qmw/ at the repository top the way /qmw:status does, and open with where the work already stands: what a past audit already found, …

    line 20
  2. Very specific:

    1. Audit the codebase: Run /qmw:audit-codebase. Present its report.

    line 39
  3. Very specific:

    3. Upgrade dependencies: Run /qmw:upgrade-deps. Present the priced upgrade list.

    line 51
  4. Very specific:

    5. Check the release: Run /qmw:check-release: clean clone, the real artifact, migrations from the last released state, the history read, …

    line 62
  5. Distinctive:

    when anything was fixed after the check ran, the check-release re-run as the final line, because the check that approves the release must …

    line 79

≈ “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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

A full cycle takes a codebase through the whole review in one pass, not one repair: audited, fixed on your orders, and checked before it ships. This skill sequences the others through that cycle. It orchestrates and gates; it does not do the work itself. Each leg is the matching skill, run in full, with its own standard of proof. Between every leg the passage stops for the user, and silence is not approval: an unanswered gate ends the passage, it does not pass it.

Work in the language of the conversation.

Before the first leg

Read .qmw/ at the repository top the way /qmw:status does, and open with where the work already stands: what a past audit already found, the refactors and dependency raises already landed, the bugs still open, the last release verdict. A full cycle often resumes rather than starts, and a leg whose work is already done is offered as skippable, not run again blind.

Then state the passage in one block: the legs below, which the project has substrate for (no front end, no run-qa leg; no dependencies in scope, no upgrade-deps leg), and which the user wants in this cycle. A user who named a narrower scope in their request skips the ones they left out. Nothing is written to the repository by this skill itself; each leg's own skill writes what it writes.

The legs, in order

Each leg hands off to its skill, which runs to its own close, including its own setup interview on a first run. This skill adds the gate between legs, nothing more, and never softens a skill's discipline to keep the passage moving.

1. Audit the codebase. Run /qmw:audit-codebase. Present its report. Gate: the user picks the findings to repair, in what order, and what they can live with. The audit repairs nothing, and neither does this leg.

2. Refactor, one finding at a time. For each finding the user picked, run /qmw:refactor on that finding alone: behavior pinned before anything moves, the class closed rather than the sites patched, proof that nothing observable changed, one commit, one entry in .qmw/refactor/log.md. A bug found mid-repair is handed to leg 5, not corrected in a restructuring commit. Gate: after each finding, report what landed and stop before the next; the user may reorder, drop the rest, or leave the cycle here.

3. Upgrade dependencies. Run /qmw:upgrade-deps. Present the priced upgrade list. Gate: the user sets the scope, everything or advisories only or named dependencies. Each raise proves out or is reverted and held with its price written down; every commit left behind is a tree that ships.

4. Run QA. Front-end products only. Run /qmw:run-qa on a genuinely empty environment, playing each persona's intentions until they land. Any bug a refactor deferred here is a report this leg carries into the pass. Gate: the pass runs continuously and only a genuine product call stops it; the passage waits on those before moving on.

5. Check the release. Run /qmw:check-release: clean clone, the real artifact, migrations from the last released state, the history read, the built product smoked the way production runs it. It fixes nothing and ends on a go or a no-go bound to one commit.

A build in flight (/qmw:build-feature) is not a leg of this passage: the cycle is the review of a codebase as it stands, not the construction of new behavior. A feature built between haul-outs is proven by its own skill and shows up here as part of what the next audit and run-qa cover.

Close

The last thing printed is the verdict of the passage and the actions, written to be pasted, nothing after them:

  • the check-release go or no-go, with the SHA it binds
  • each blocker as the follow-up that clears it
  • when anything was fixed after the check ran, the check-release re-run as the final line, because the check that approves the release must be the one that saw it

A cycle that reached a no-go is not a failure of the passage: it is the check doing its work before the release ships instead of after. Report it as plainly as a go.

Say where the passage broke its own shape

One line per deviation, before the closing actions, only when there is one: a leg run without its gate, a skill's discipline softened to keep moving, a leg silently skipped rather than named as skipped. Nothing to say is the normal case, and then print nothing at all. A passage that quietly ran a leg past its gate reports a consent the user never gave.

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.