agentleFS
AlexaysAlexays/Lilo/.claude/skills/changelog-draft/SKILL.md

changelog-draft

Skill58 starsChanged 10 days agoMITView on GitHub

Draft a Lilo release's changelog entry (website story, What's New highlights and their copy in every catalog, media plans) from its release notes. Use when asked to draft, write or update the changelog entry, release story or What's New highlights of a release.

4 other projects use a version of this.

Adopted by

Most recently taken up by

  • Other versions4 projects use a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 4

    1. warpdotdev/Warp★ 66ksame name
    2. warpdotdev/warp★ 66ksame name
    3. dominikmartn/harp★ 715same name
    4. black12-ag/weft★ 2same 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 repository6 other files it gives its agents.

The file

We split this skill into the rules it gives an agent

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

  • 0

    Common

    Most of the 4 edited versions we read keep these.

  • 0

    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
  1. Very specific:

    Write the draft to build/changelog/<version>.json in the shape below.

    line 13
  2. Very specific:

    Run make changelog-draft DRAFT=build/changelog/<version>.json, with the same RELEASE.

    line 14
  3. Very specific:

    Every highlight shows a visual in What's New: its illustration always, replaced by a film when its plan is captured.

    line 45
  4. Very specific:

    Use Lilo's own terms as the glossary spells them: Space, Private Space, Agents Space, pinned tab, favorite, folder, archive, Peek, Little …

    line 74

“Common” is measured: kept by at least 6 in 10 of the 4 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

Draft a changelog entry

You write the changelog of Lilo, a native macOS web browser built on WebKit. Readers are people who use Lilo every day, not developers of it. make changelog-draft checks your draft and writes every file from it; you never edit the story, the catalogs or releaseHighlights.json by hand.

Steps

  1. Run make changelog-context, with RELEASE=x.y.z for a published release (pending notes otherwise). It prints the numbered notes, the three latest entries as examples, the commits, the interface diff and the locales to translate into. Stop if it says the release already has an entry or has only fixes.
  2. Write the draft to build/changelog/<version>.json in the shape below.
  3. Run make changelog-draft DRAFT=build/changelog/<version>.json, with the same RELEASE. When it refuses the draft, fix the draft and run it again: nothing is written until every check passes.
  4. Show the person the English and French story and wait for their review. Commit, push or open a pull request only when they ask. On a changelog/<version> branch, the Changelog media workflow films the highlights; locally, make changelog-media RELEASE=x.y.z does.

Content

Write the entry from the release notes, using the commits only to understand what a note means. Every claim must come from a note: never invent a behavior, a number, a shortcut or a benefit. Commits can explain a note, but they never add a claim of their own.

{
  "title": { "en": "...", "fr": "..." },
  "introduction": { "en": "...", "fr": "..." },
  "highlights": [
    { "id": "...", "title": { "en": "...", "fr": "..." }, "body": { "en": "...", "fr": "..." }, "notes": [0], "illustration": "...", "plan": null }
  ],
  "groups": [{ "title": { "en": "...", "fr": "..." }, "notes": [1, 2] }],
  "translations": {
    "de": { "headline": "...", "description": "...", "highlights": [{ "title": "...", "description": "..." }] }
  }
}
  • title: a short headline in the voice of the example entries, three to six words, ending with a period. It names the release's theme, not a list of features.
  • introduction: one sentence on what the reader gets, in plain words.
  • highlights: one to three features worth a section, the ones a reader would tell a friend about. An id is a short kebab-case word. title is three to six words and ends with a period. body is one or two sentences, 45 words at most: open with what the reader can now do, then say how, naming the actual control or shortcut.
  • groups: every note that is not a highlight, gathered under short area names such as "Sidebar", "Calls", "Agents" or "Performance". A group's title has no period. Use three to seven groups, ordered by how much a reader cares. A note goes in exactly one group. Fixes go with their area, not in a group of their own.
  • translations: the English title, introduction and each highlight's title and body, translated into every locale make changelog-context lists, highlights in the same order. Natural, calm first-party product language, keeping the same meaning and product terms.

Refer to notes by their number in notes. A highlight lists the notes it tells; those notes then stay out of groups.

Media

Every highlight shows a visual in What's New: its illustration always, replaced by a film when its plan is captured. Weigh the visual when picking highlights: between two features of equal interest, take the one the interface shows.

illustration names a drawing in lilo-ui/src/components/whatsnew/WhatsNewArt.svelte. Reuse one when it fits the feature; otherwise draw a new one before running make changelog-draft: add its identifier to illustrations in releaseNotes.ts and a {:else if kind === "..."} branch in WhatsNewArt.svelte, in the style of the existing ones (a 128 by 90 calm, decorative sketch built from the interface's color tokens, no text). The tool refuses a highlight whose illustration is not drawn.

A highlight also gets a plan when a picture says more than its words: it drives the real app and films the change. Leave plan null for anything the interface does not show.

A plan is { "template": "...", "steps": [...] }. Pick the template that fits:

  • image: one finished state. One shot.
  • hero: one state worth a slow look. One shot.
  • compare: a before and an after. Two shots.
  • flow: an interaction, step by step. Two or more shots, with a click between them; a pointer travels to each click.

A step is { "do": "...", "selector": null, "text": null, "value": null, "name": null }, unused fields null. Steps run in order in a fresh window with one Space, a sidebar on the left and no tabs.

  • page: open a sample example.com page, with text as its title. Open at least one first.
  • call: open a sample call, which shows the sidebar's call card.
  • action: run the native command name: openSettings, openHistory, editCurrentSpace, pinCurrentTab, newTab or toggleSidebar.
  • click: click the selector, narrowed to the element whose text is text when set.
  • input: type value in the field at selector.
  • wait: wait for selector to appear. Use it after anything that opens.
  • shot: capture selector, as tight as the change allows. Close-ups read well and whole windows do not: leave selector null only when the whole window is the point.

Use only class names that appear in the interface diff, and only reach for what you can see there. The steps cannot drag, hover or use a keyboard shortcut: when a change needs one, leave plan null. Sample content is generic and invented, never a real site or person.

Voice

Calm, plain and concrete, like a first-party Apple release note. Prefer "you" and the present tense. Show the benefit before the mechanism. Name interface items exactly as the notes do. No hype, no exclamation marks, no "we", no "now you can finally", no rhetorical questions, and no em dashes.

Use Lilo's own terms as the glossary spells them: Space, Private Space, Agents Space, pinned tab, favorite, folder, archive, Peek, Little window, Split View, command bar.

French is written, not translated: natural French with the same meaning, using "vous", with French typography (a non-breaking space before : ? !). Keep the product terms Lilo's French interface uses in the example entries.

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.