agentleFS
blauwtjeblauwtje/exo/skills/document/SKILL.md

document

Skill3 starsChanged 57 days agoMITView on GitHub

Writes and edits prose read later: README, doc page, PR body, changelog line, release notes. Use when the user asks for one. Not for chat replies or code comments.

56 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use develop, scope and 9 more.

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.
  • hong9LeeSkill

    develop

    빌드, 아키텍처, git 관련 작업 시 OMS MSA 공통 인프라 규칙을 적용한다. 기술 스택, 빌드 명령, 헥사고날 아키텍처 구조, git 워크플로우, AI Context 동기화 규칙을 다룬다. 코드 작성 컨벤션은 /convention, 테스트 컨벤션은 /test-guide 스킬을 참조.

    Also in 5 projects that use this

  • YuriNachosSkill

    scope

    Configure claude-scope status line widgets for Claude Code. Use when user wants to customize widgets, change theme, reorder elements, or modify display style.

    Also in 5 projects that use this

  • ralphkraussSkill

    resolve-conflicts

    Resolve git merge conflicts or re-apply changes from a stale branch onto current main. Use when user says "resolve conflicts", "merge conflicts", "rebase branch", "branch is behind", "re-apply changes", "port branch", or a branch is too far behind main to merge cleanly.

    Also in 3 projects that use this

  • MicrosoftSkill

    sync

    Sync the current session branch with its upstream branch, or publish the current session branch to a remote. Use when the user asks to sync a branch, pull latest changes, rebase onto upstream, push current branch, publish branch, or set upstream.

    Also in 5 projects that use this

  • kt057Skill

    check

    lint, typecheck, test を一括実行して結果をサマリーする

    Also in 7 projects that use this

  • Microsoft AzureSkill

    test

    Also in 12 projects that use this

The file

We split this skill into the rules it gives an agent

6 rules. 0 are common, 4 are distinctive, and 2 are very specific to this project.

  • 0

    Common

    Most of the 52 edited versions we read keep these.

  • 4

    Distinctive

    This author's own choices. Worth a look.

  • 2

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Name the actor and the mechanism: Use the active voice with the real file, option, flag and command names, and one name per thing.

    line 17
  2. Distinctive:

    Change only what the request names: Run the doc's link or reference check when one exists.

    line 19
  3. Distinctive:

    Scan the draft for the tics below and rewrite each hit: When reviewing someone else's text, quote each hit.

    line 20
  4. Distinctive:

    When the user asks for release notes or the changelog lines a range misses:

    line 34
  5. Very specific:

    Run node "${CLAUDE_SKILL_DIR}/scripts/release-notes-draft.mjs" --repo <checkout> [--from <ref>] [--to <ref>].

    line 36
  6. Very specific:

    Add --uncovered to list only commits that did not touch the changelog.

    line 38

“Common” is measured: kept by at least 6 in 10 of the 52 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

Document

Prose for a reader who was not in the session says who does what, by which mechanism, in the plain word. Avoid text that sounds finished and says little, and avoid the opposite, telegraphic text without articles and verbs that the reader must decode.

Process

  1. Pick one mode per document. A tutorial teaches by doing, a how-to gets a known task done, a reference lists what exists, an explanation says why. A how-to opens on the command, with no welcome or tour.
  2. State the fact, not the benefit. Give the number, behavior or error that changed. When asked to sell or impress, still write only what the input states, because an unmeasured effect is invented.
  3. Name the actor and the mechanism. Use the active voice with the real file, option, flag and command names, and one name per thing.
  4. Write whole sentences. One thought per sentence, under 25 words. Put the condition before the instruction and the warning before the step it guards. Put "only" and "not" next to the word they limit.
  5. Change only what the request names. Run the doc's link or reference check when one exists.
  6. Scan the draft for the tics below and rewrite each hit. When reviewing someone else's text, quote each hit.

Tics

  • Trailing "-ing" benefit phrase ("ensuring stable memory"), inflated significance, a vague source, a promise about the future. State the number or cut it.
  • Quality adjectives with no measurement (robust, seamless, significantly). Give the measured change.
  • Stock vocabulary (delve, pivotal, landscape, showcase, leverage, utilize, facilitate), "serves as" for "is", "not just X but Y", lists of three for rhythm, and "simply" or "just".
  • Em dashes, a colon for drama, bold on ordinary words, Title Case headings, emoji, curly quotes around code, headings on a one-paragraph text.
  • Chat traces: "Here is...", praise, "I hope this helps", "That's it!", a section that describes itself, a summary repeating the list above it.
  • Hedges stacked ("could potentially"), filler ("in order to", "it is worth noting"), aphorisms, abstract metaphors (backbone, guardrail), and "feels smoother" where a mechanism belongs.
  • A passive verb with no actor, a pronoun with two possible referents, a noun stack, a semicolon joining sentences.

Release notes

When the user asks for release notes or the changelog lines a range misses:

  1. Run node "${CLAUDE_SKILL_DIR}/scripts/release-notes-draft.mjs" --repo <checkout> [--from <ref>] [--to <ref>]. --from defaults to the last v* tag reachable from --to, --to to HEAD. It prints ### Breaking changes, ### Added, ### Changed, ### Fixed and ### Removed from non-merge commits, and adds merged PR titles through gh unless --prs <file> gives a JSON list of { number, title, mergedAt }.
  2. Rewrite each line to say what the user sees. Fill each Upgrade: not stated in the commit. line from the commit diff, and keep the Upgrade: line under each breaking entry.
  3. Add --uncovered to list only commits that did not touch the changelog; in the changelog a breaking change stays in its type section and starts with Breaking:.
  4. Name the range used in one line, so the user can rerun it. Edit the changelog only when the user names it, and never tag, bump or release.

Judgment

  • The codebase's name outranks the plain word: maxBytes stays maxBytes.
  • A fact, number or term the reader needs outranks every tic fix, and a tic inside a quotation, command or code sample stays as written.
  • The pull request body layout of exo:ship and the brief sections of exo:spec outrank this skill's layout, and product interface strings follow the product's copy rules.

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.