agentleFS
FracteraFractera/web4-node-starter-kit/.claude/skills/use-use-cases/SKILL.md

use-use-cases

Skill68 starsChanged 2 days agoMITView on GitHub

Mode `cases` — the order in which work is allowed to start when the owner has switched it on: a case he confirmed, then a queue of numbered steps, and every step naming the case it serves. Load it when `PLATFORM-CONFIG.mode` is `cases`, when the owner says "add a case" / "what are we building next" / "decompose this", and before opening any step in a product that has a dossier. Two things you cannot guess: a case is confirmed ONLY by the owner and nothing in your reach can confirm one, and the decomposition step is found by its `kind`, never by matching its title — the title is a sentence somebody will reword.

2 other projects use a version of this.

Adopted by

Most recently taken up by

The file

Files
SKILL.md

use-use-cases

Informational, not binding. Know a better way for the case in front of you — do it your way and say so. You are trusted with the creative work on this project.

The mode answers one question: what gives you the right to write the first line? In classic the answer is "he asked". Here it is a confirmed case, and everything below is the machinery that keeps that honest.


1. Where the two entities actually live

CasesSteps
What it iscontent a person confirmedthe state of a queue
Homethe product dossier — PRODUCTS-CONFIG/<id>.json, array cases[]table development_steps (declared in lib/db/index.ts), mirrored in the dossier's steps[]
A case carriesslug, title, summary, confirmed, confirmedAt, updatedAtnumber, product_id, title, status, importance, kind, cases, plan, result
Asked of it"what are we building""what is open for this product"

🔒 The step NUMBER is server-wide and permanent. Not per product: the dossier lists a product's steps by number as a table of contents, and a number that only means something inside one product would make that list unreadable. Closing a step never renames or renumbers it — completion is a column (status), not a rename.

🔒 product_id = 'platform' is a legal value, not a placeholder. Theme, languages, the offline cache belong to the whole server; forcing a product id onto them would be a field that lies.

2. The flow, and why it has two entrances

intake — the panel's Quiz writes questions and answers into the dossier
      │
      │  🔒 the OWNER confirms a case  (confirmed: true)
      ▼
a DECOMPOSITION step appears           kind: "decomposition", one per product
      │  created by the panel at confirmation time;
      │  created by the agent at session start if it is missing
      ▼
you execute it: read the confirmed cases, write the ordered queue
      the FIRST step is always the minimal working skeleton —
      architecture on disk, doors present, navigation walking on stubs
      ▼
development → analysis (the owner decides what is finished)

Two independent paths to one state, both idempotent. The panel creates the step because that is where the owner acts and the queue must exist even if no agent is ever opened. You create it because a mechanism that depends on the other path having run is a mechanism that silently does nothing.

🔒 Find it by kind, never by title. WHERE product_id = ? AND kind = 'decomposition'. A lookup that matches a sentence survives exactly until somebody rewords the sentence.

3. 🔒 Rules that decide whether the mode means anything

  • Only the owner confirms. There is no tool to confirm a case — by construction, not by omission. An unconfirmed case is the model's guess about his business, and building on it is the expensive kind of wrong: it looks finished.
  • Editing a confirmed case returns it to draft. Green must mean "he approved this text", not "he approved something that used to be here". The panel does this; do not work around it.
  • A step names the cases it serves (cases: string[]). A step serving none is work nobody ordered. platform steps are the one exception.
  • A step's title is 6–12 words — the service refuses anything else. A title is a promise, not a paragraph.
  • The stage only moves forward. Backwards is the owner's decision, made in the panel.
  • The phase is derived, not typed: intake → decomposition → development → analysis come out of the dossier itself — the questions answered, the cases confirmed, the statuses of the steps — so it cannot drift from the truth. Details of the dossier: use-products-config.

4. 🔒 The closing question is not "does it work"

Green gates and a live page answer "does it work". They cannot answer "is this what was ordered" — only the case can. So closing a step, name its case and say what promised in it is now true. A feature that works perfectly and answers no case is work the owner did not ask for, and it is the most expensive thing this mode exists to prevent.

5. The boundary with the other two skills

SkillOwns
use-use-cases (this)the right to start: cases, their confirmation, the queue born from them
build-product-with-ownerthe PATH once a case is confirmed: the pact with the owner, the four decisions before the first route, the furniture, the size of an iteration, the prototype line
use-development-stepshow a step is CARRIED — session handover in current-steps.md, the group, the closing report
use-products-configthe dossier itself: record, phases, the pages plan, what belongs to which product

They are not alternatives. The table is the queue the owner sees in the panel; current-steps.md is what survives a context reset. Neither replaces the other, and neither is derived from the other.

6. Before you say the mode is being followed

  1. The case you are building against is confirmed: true in the dossier — quote its slug.
  2. The step exists in the queue and names that slug.
  3. The first step of a new product is a skeleton that RUNS, not a folder of intentions.
  4. Nothing was confirmed, moved backwards, or reworded by you on the owner's behalf.

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.