agentleFS
InditexTechInditexTech/weavejs/.agents/skills/docouture-getting-started/SKILL.md

docouture-getting-started

Skill246 starsChanged 46 days agoApache-2.0View on GitHub

How to take a repo from no docs at all to a real, published docouture site — scaffolding it with the docouture CLI if it isn't already, planning what the site's structure should be, and drafting first content from whatever the repo actually has (existing docs, README, CONTRIBUTING, or the code itself). USE WHEN starting documentation from scratch, a repo has no docs/ yet, deciding what a new docs site should cover, or planning its module/page structure for the first time. EXAMPLES: 'set up docs for this repo', 'I want to start documenting this project', 'what should my docs site look like', 'scaffold docouture here and write a first draft'.

We split this skill into the rules it gives an agent

1 rule. About 0 are common, 0 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈0

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

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

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

    Scaffold, if it isn't already: Check for docs/antora-playbook.yml (or run docouture doctor — a missing-site error means there's nothing …

    line 39

≈ “Common” is an estimate: at least 1 in 5 of the 308,232 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

Getting started with docouture

This is the entry point for a repo with no docouture site yet. It gets you from nothing to a published-shaped site with a real first draft — not perfect, not complete, but grounded in what the repo actually is rather than a boilerplate stub. Once a site exists and is evolving normally, docouture-documenting-changes takes over for keeping it in sync as the repo grows; re-entering this skill after that point should be rare (a full restructure, not routine maintenance).

This is a planning and orchestration skill, not a syntax reference. Once a decision is made here, it hands off to three skills:

  • docouture-docs-internals — once you know a module/page needs to exist, this is how antora.yml, nav.adoc and the home-page pattern actually work.
  • docouture-authoring-guides — once you know a page needs to exist, this is what it should say: the section skeleton, per-section instructions, a copyable AsciiDoc skeleton and a quality checklist, for every page in the standard six-section structure plus the home page.
  • docouture-writing-docs-pages — once you know what a page should say, this is how to write it in AsciiDoc: the syntax, xrefs and this site's custom blocks ([tabs], [cards], [accordion], …).
  • reference/structure-planning.md — mono- vs. multi-module decision, the default nav structure, how Guides/Reference get derived from the repo rather than assumed, the per-artifact-module strategy for a monorepo, and the home page's special (structure now, content later) treatment.
  • reference/content-sourcing.md — where content comes from when nothing is prescribed: priority order across existing docs, README/CONTRIBUTING, and the repo's code itself (exports, doc-comments, CLI definitions, config schemas, tests-as-examples) — including the all-code, zero-prose case.

The flow, in short

  1. Scaffold, if it isn't already. Check for docs/antora-playbook.yml (or run docouture doctor — a missing-site error means there's nothing here yet). If absent, run docouture new <name> (interactively, or with flags — see the CLI's own --help) from the repo root. This is the CLI's whole job: it lays down the starter site and GitHub workflows, nothing about content or skills.
  2. Base check — confirm identity/branding docouture new already seeded (title, product name, description, favicon, light/dark logo). Look for existing brand assets elsewhere in the repo before asking the user to supply new ones.
  3. Structure planning — decide which modules exist beyond the default main. See reference/structure-planning.md. The home page's slot is asserted here — its content is not, see below.
  4. Content sourcing & drafting — for every planned page, find where its content actually lives (don't assume — a repo with no README at all still has an API surface, a CLI, a config schema) and draft it. See reference/content-sourcing.md. Before writing, open that page's guide in docouture-authoring-guides for its section-by-section contract (what to say, what not to say, the copyable skeleton and quality checklist) — content-sourcing decides where the facts come from, the authoring guide decides what shape the page takes. Draft a rough home page early from whatever one-line description exists; treat it as unfinished.
  5. Record what was sourced from code in AGENTS.md's documentation-state ledger — see the docouture-documenting-changes skill's reference/maintenance-loop.md for the exact shape. This is what lets that skill pick up the loop from here, later.
  6. Hand off. Once the site has a real first draft, day-to-day upkeep belongs to docouture-documenting-changes — don't keep re-running this skill's structure-planning pass every time something changes; that skill's re-entry loop is for that.

What this skill does not opinionate about

Where in the repo the "real" information lives is entirely up to the repo itself — a doc-comment, a type signature, a schema file, a test, a --help string, an existing README.md, are all fair game and equally valid sources. This skill does not prescribe a required location or format for source material; it only prescribes that every documented surface is traceable to something real, and where in AGENTS.md that traceability is recorded.

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.