본문으로 건너뛰기
cccskills
무료GitHub에 공개한국어 소개

ui-copy

사용자가 상황과 다음 행동을 이해할 수 있도록 영어 화면 문구를 작성하고 검토합니다. 버튼, 오류, 설정, 알림, 접근성 레이블의 용어와 표기를 일관되게 맞추며, 기존 문구의 의도된 어조를 보존하는 기준도 제공합니다.

원문원문 설명 보기

Copy rules for every user-visible string in riptide-ui, riptide-cloud, packages/ui, and server strings that reach a toast. ACTIVATE when adding, changing, or reviewing a button, menu item, dialog, toast, error, empty state, placeholder, tooltip, aria-label, setting, hotkey description, banner, or onboarding text. Not for apps/docs, the CLI, or comments.

humanlayer/humanlayer스타 12,707(GitHub 12,707 + cccskills 0)
연동 도구Jira

실행 점수 미검증

설치 방법 보기

이럴 때 유용합니다

  • 명확한 버튼 및 메뉴 문구 작성
  • 오류와 다음 조치 안내 개선
  • 용어와 표기 일관성 확인
  • 빈 화면과 스크린 리더 레이블 검토

한국어 소개

할 수 있는 일

버튼, 메뉴, 오류, 알림, 설정 등 화면에 표시되는 영어 문구를 작성하고 검토합니다. 시스템 내부 상태를 설명하는 데 그치지 않고, 사용자에게 어떤 의미인지와 다음에 무엇을 할 수 있는지를 전달하도록 다듬습니다. 대소문자, 문장부호, 숫자, 키 이름, 제품 용어, 스크린 리더용 레이블에 관한 구체적인 규칙도 제공합니다.

이럴 때 유용합니다

화면마다 용어가 다르거나 오류를 읽어도 다음 조치를 알기 어려울 때 유용합니다. 기존 문구를 검토할 때는 기계적으로 고칠 수 있는 오류와 어조에 관한 판단을 구분하고, 필요한 부분만 수정합니다. 의도적으로 개성을 담은 표현을 일괄적으로 바꾸지 않도록 안내합니다.

사용 예시

  • “이 오류가 작업에 미치는 영향과 다음 조치가 드러나도록 고쳐 줘.”
  • “버튼과 설정 항목의 표현이 일관적인지 확인해 줘.”
  • “빈 목록 안내와 스크린 리더용 레이블을 검토해 줘.”

주의 사항

대상은 riptide-ui, riptide-cloud, packages/ui, 그리고 알림에 표시되는 서버 측 문구입니다. apps/docs, CLI, 주석은 대상이 아닙니다. HumanLayer 전용 용어집이 포함됩니다. 홍보 문구와 주요 제목은 사실 오류가 없는 한 검토 대상에서 제외됩니다.

이 소개문은 공개된 SKILL.md를 바탕으로 AI(OpenAI Codex)가 작성했습니다. 정확한 사양은 아래 원문을 확인하세요.

문서 점검

AI가 SKILL.md의 내용에서 주의가 필요한 작업과 도입 전제를 확인했습니다. 점수는 실제로 실행한 결과(실행 검증)만으로 매깁니다.

안전성전송·삭제·명령 실행 등을 명시하고 있는가
A양호문구 작성과 검토를 다루며 외부 전송이나 파괴적 작업은 설명하지 않습니다.
도입의 용이성필요한 도구나 API 키가 적고 적혀 있는가
B일부 부족대상 저장소와 용어집은 명시하지만 설치 단계나 요구 사항은 없습니다.

점수에는 넣지 않았습니다. 2026년 10월 29일 기준으로 AI(OpenAI Codex)가 판정했습니다. 참고용이므로 설치 전에 원문도 확인해 주세요.

평가가 다르다고 생각되면 재검토를 요청하기

어느 항목이 왜 다르다고 생각하는지 적어 주세요. 로그인은 필요 없습니다. SKILL.md와 대조해 평가를 다시 만들고 이곳에 반영합니다.

이 양식은 reCAPTCHA로 보호되며 Google 개인정보처리방침 및 서비스 약관이 적용됩니다.

포함된 파일(1)

  • SKILL.md14.1 KB

SKILL.md(원문)

설치하기 전에 에이전트에게 전달되는 지시 내용을 확인할 수 있습니다.

First rule: write to the person, never about the system. Every string answers "what does this mean for me and what can I do now?", never "what state is the code in?". A string that narrates internal state, preconditions, or implementation progress is a log line, not copy, even when it is accurate. Before any rule below, reread the string as the user and ask what they would do next. Workspace setup is running independently. No session has started. → Setting up your workspace. Your session starts when it finishes. Worktree configuration is required to use "Now". → Add a workspace config to create worktrees now. Task creation sync timed out → Could not confirm the task was created. Check the task list before retrying. Comment sync still loading → Comments are loading. Workspace setup row is parked on an error → Workspace setup failed. Converging, Delivery rejected: … → never.

Voice: a competent colleague at the next desk. Says the actual thing, names the key, admits limits. Never apologizes, sells, or shouts. Model string: Updating will interrupt 3 active sessions on this device. Finish your work before updating, or choose Later.

Changing a string that already ships

Writing a new string: apply everything below. Changing one that exists: read this first. A rule matching a line is not on its own a reason to rewrite it.

  • Two kinds of finding. Mechanical — a raw enum, Title Case, ..., ENTER, (s), an internal name, Failed to, "please", a dead route. One right answer, no taste, fix it. Voice — hero lines, warnings, tips, empty states, anything with a deliberate point of view. Here the rule is a prompt to look, not a verdict.
  • Change the defect, not the sentence. If one word is wrong, fix that word. …how Riptide queries data… You probably don't want to mess with these. needed Riptide → HumanLayer and nothing else.
  • If the rewrite is only more conformant, stop. It must be clearer, more honest, or more useful to the reader. "Now it follows the guide" is not a benefit anyone experiences.
  • Never bulk-rewrite voice. Mechanical findings batch; voice findings go one at a time, with the tension named, for a person to decide. An audit that proposes 180 rewrites has stopped reading and started matching.
  • Hero and marketing copy is out of scope unless it is factually wrong. The install page and the composer headline are written on purpose.
  • When you are unsure which kind you have, it is voice. Leave it and say why you looked.

Mechanics

  • Sentence case everywhere: buttons, titles, headings, menu items, labels, badges, tabs, toasts.
  • Proper nouns keep casing: HumanLayer, HumanLayer Cloud, Claude Code, Codex, Fold, GitHub, Linear, Jira, Slack, Stripe, macOS, Starter/Pro/Enterprise, RPI, PRD, TDD.
  • Feature names are common nouns: fast mode, bypass permissions, auto-accept edits, multiplayer prompting, advanced metrics, commit tracking.
  • Never uppercase in the source string: write Needs approval, not NEEDS_APPROVAL; DENY → Deny. Rendering is then a CSS decision in one place, and the default is not to uppercase. The terminal-style caps on badges and eyebrows accreted rather than being designed, and the app reads better without them; keep them only where a surface earns it.
  • Never render a raw enum (needs_approval, xhigh). Add a label map first.
  • Full sentences end with a period. Fragments (buttons, labels, badges, placeholders, toast titles, headings, hotkey descriptions) do not.
  • Ellipsis is …, only on in-progress labels: Saving…. Never ..., never on placeholders.
  • Em dash with spaces for a consequence. · between status facts (This device · online). → for sequences. – for ranges (2–4×). Oxford comma. Straight quotes.
  • No !. & only in a compact two-verb button (Interrupt & send); prose uses "and". No + in prose.
  • Address the user as "you".
  • Actor: you do what you click (install the GitHub App, connect Linear, start a daemon). The agent does the work in a session (fixes, implements, opens a pull request). HumanLayer only does what happens on its own with no click (syncs, checks this device, sends a notification). Otherwise no subject: Could not save draft. Connect GitHub so the agent can open pull requests. HumanLayer checks this device before your first task.
  • Never the company "we" — a faceless narrator reporting or apologising: We couldn't fetch your organizations, Something went wrong on our end, we'll investigate. Rewrite with no subject or with HumanLayer as the subject.
  • The collaborative "we" — you and the agent, as a pair — is allowed, and only in the composer where that pairing is literally what is happening: What should we build today? stays. Do not extend it to status, errors, or settings.
  • Never "please", "sorry", "successfully", "just", "simply", "Something went wrong", "Are you sure". No roadmap hedges: "currently", "not yet", "coming soon". yet in an empty-state fact (No tasks yet) is fine.
  • Do not sand off personality. A line with a point of view is not a defect. You probably don't want to mess with these is doing a real job; flattening it to Change these only when asked to is the machine talking. Fix what is wrong in a line (an internal name, a raw enum) and leave the rest alone. If a rewrite is only blander, do not make it.
  • Use contractions (can't, don't, isn't). The one exception is cannot in irreversibility warnings: This action cannot be undone.
  • No internal names (Riptide, CodeLayer, Launcher V3, Projects V2, Phase 5, route paths, ISO timestamps, WorkOS).
  • Digits, real plurals (1 session, 2 sessions, never (s)), thousands separators. Zero is a word: No sessions, not 0 sessions.
  • Time: relative in lists (3 min ago, yesterday), absolute in the tooltip (Sep 10, 2026, 15:04), durations as 2m 3s.
  • Truncate the end of a title and the middle of a path (~/code/…/riptide-ui). Never truncate a number or a key.
  • Keys: ⌘ ⇧ ⌥ / Ctrl Shift Alt, joined with +, Enter, Esc, uppercase letters. Always give the non-mac form. Never Cmd, ENTER, ↵, ⌃. A key is a keycap element next to the label, never typed into the label string. In this document a key shown after a label means a keycap.

Session status labels

valuelabel
draftDraft
ready_for_launchQueued
waiting_for_workspacePreparing workspace
launching, resumingStarting
runningRunning
ready_for_inputIdle
needs_approvalNeeds approval
interrupt_requestedInterrupting
interruptedInterrupted
failedFailed
lostStalled

Shapes

typeshapeexample
Button, menu itemverb + object; bare verb when the object is the dialog subject; in-flight = gerund + …Create task, Delete, Creating…. Not OK, Yes, Submit, Confirm, Got it
Success toasttitle = object + past participle; partial success: title names the win, description the missDraft saved. Token created / The token list could not be refreshed.
Error1 what happened Could not <verb> <object> 2 what it means for your work 3 what to do or an action buttonCould not send. Your draft was restored. Not Failed to, Unable to, Error, Unknown error
Blocked statethe precondition as an instructionConnect a host to launch
Integration promptConnect <Product> to <payoff>; payoff is one clause naming what appears or what the agent can do; optional cost hint as status factsConnect GitHub to see pull requests and checks here · about a minute · opens github.com
Capability limitlimit + alternativeTerminal is not available for sessions on remote hosts.
Permission denialwho can fix itOnly organization admins can manage billing. Ask an admin.
Retryonly when it can succeed, with a how or when; never bare Please try again.Refresh billing and try again.
Raw backend textnever a title; description or Show details
Confirm dialogtitle <Verb> <object>?; body = consequence, what is preserved, undoable?; confirm = title verb; Cancel + Esc; confirm = ⌘+EnterArchive project? / Existing workspaces are not deleted. / Archive
Reversible actionno dialog; undo toastSession archived + Undo Z
Warninglist the consequences; never an intensifiernot Use with extreme caution!
Empty statefact, then what fills it or what to do, with keyNo tasks yet / Create a task to organize your sessions + Create task T
Loadingname the work or actorCreating worktrees…, Asking the agent…; plain Loading… only when nothing more is true
Step listshort noun phrase; states Pending / In progress / Complete / Failed / SkippedPrepare launch
Placeholdernoun phrase, instruction, or format example; no …, no period; a key hint is keycap + textSearch tasks, you@company.com, ENG-123 or a Linear issue URL, Enter to start typing
Setting labelnoun phrase, not Enable XDefault coding agent
Setting descriptionone sentence: on, and off if not the obvious inverse; experiments say what changes and what does notSend all queued messages together as soon as the session is ready. Turn off to send one at a time.
Org lockYour organization has turned this off.
Tooltip (icon button)action label + keycapToggle terminal ⌘+T
Tooltip (explainer)one or two full sentences
aria-label (control)verb + object, no key, no periodCopy file link
aria-label (region)noun phraseDocument outline
Toggle state (aria){label}: on / : off
Hotkey descriptionverb + object, no articleOpen command palette. Reuse hotkeysCatalog.ts wording
Bannertitle fragment + one-sentence body; route non-admins to a personPayment overdue / Update billing details to avoid blocked writes.
Notification toasttitle = event; body = session or task; action = verb + keyNeeds approval / Jump to session ⌘+⇧+J
Money, ratesas charged$3.84 / commit, 2–4×

Glossary

conceptusenot
productHumanLayer; HumanLayer CloudRiptide, CodeLayer, Codelayer, the app, the application
unit of worktaskticket (= the Linear/Jira issue), project
one agent runsessionconversation, chat, run
unlaunched task/sessiondraft, Untitled draftUntitled session
reusable repo recipeprojectworkspace, recipe
session output documentartifactdocument, doc, file
repo file or attachmentfileartifact
per-task checkoutworkspace, workspace setupworktree setup
git worktreeworktreeworkspace
top-level panetabworkspace
machine that runs sessionshost; this device; remote hostcomputer, machine, worker, daemon host, locally
process on a hostdaemon (only when starting, stopping, installing)host
the AI in prosethe agent; coding agent (category); subagentClaude, bot, model, sub-agent, assistant
the AI as an identity label (speaker, badge, picker)the product name: Claude Code, Codex, Foldclaude code, codelayer
tenantorganizationorg, team, workspace
people in itmember; teammates when invitinguser
rolesadmin, member; owner (of a task/artifact)administrator, org admin
tool-call decisionapprove / deny; allow (standing rule)accept, reject
hide reversiblyarchive / unarchiveun-archive, remove
to trashdeletearchive, dismiss
close a nudgedismissdelete, hide
thread lifecycleresolve / reopenmark as resolved, close
comments to agentsend to agent; send and resolvecopy for agent (= clipboard)
permission modesbypass permissions, auto-accept edits, manualYOLO, Auto-Accept, default
priority tierfast modeFast Mode, priority-tier
reasoning budgeteffortreasoning effort
GitHub PRpull request (prose), PR (labels)merge request
metricsTeam metrics, Advanced metrics, cost per commit, cost per 1,000 lines changedkLOC
filesystem locationdirectory; path (the string)folder

Open product decisions, do not add a third name: the codelayer agent option is labelled Codex, shows as CodeLayer in badges, and becomes Fold under a flag; workflows have two label sets (RPI / PRD / TDD / Freeform / Exploratory and Small / Medium / Large / Exploratory).

Rewrites

beforeafter
Failed to approveCould not approve the tool call. The agent is still waiting.
We couldn't fetch your organizations. Please check your connection and try again.Could not load organizations. Check your connection and retry.
Something went wrong on our end. Please try again later.Could not reach HumanLayer. Try again in a moment.
An unexpected error occurred. The error has been reported and we'll investigate the issue.This page could not be rendered. The error has been reported. + Try again
User role updated successfully / Draft saved!Role updated / Draft saved
APPROVE & DON'T ASK AGAINApprove and don't ask again
Use with extreme caution!The agent will edit files, run commands, and call tools without asking. Turn it off any time with ⌥+Y
ENTER to start typing...Enter to start typing
No hosts available. Please start a daemon first.No hosts online. Start a daemon to launch sessions.
Organization created, but 2 invite(s) failed. Manage invitations at /admin/invitationsOrganization created, but 2 invitations were not sent. + Manage invitations
Agent delivery is added in Phase 5remove the control until it ships

리뷰

아직 리뷰가 없습니다. 사용해 본 소감을 남겨 주세요.

댓글

로그인 없이 쓸 수 있습니다. 사용 후기, 질문, 작동하지 않은 환경 등을 편하게 남겨 주세요.

아직 댓글이 없습니다. 첫 댓글을 남겨 보세요.

보내면 바로 공개됩니다. 이메일 주소나 전화번호 같은 개인 정보는 쓰지 마세요. URL은 2개까지입니다.

0 / 1,000자

이 양식은 reCAPTCHA로 보호되며 Google 개인정보처리방침 및 서비스 약관이 적용됩니다.

같은 저장소의 스킬

Automate website navigation, form entry, and clicks through a browser command line tool. Extract page text, capture screenshots, and compare page states to check interactions or test web app behavior across screen sizes.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

  • Automate form entry and button clicks
  • Extract text from web pages
  • Save page screenshots and PDFs
humanlayer/humanlayer1.3만2026년 10월 29일 업데이트

asd-ste100

무료요약 있음

ASD-STE100의 원칙을 바탕으로 모호하거나 빽빽한 영어를 짧고 직접적인 문장으로 고칩니다. 사실과 조건을 보존하고, 문장 문제와 수정 이유를 설명하며, 원문과 수정문을 표로 비교합니다.

  • 다른 에이전트에게 줄 영어 지시 명확히 하기
  • 영어 도구 설명이나 오류 문구 간결하게 만들기
  • 영어가 모국어가 아닌 독자를 위한 문장 정리
humanlayer/humanlayer1.3만2026년 10월 29일 업데이트

better-github-skill

무료요약 있음

전용 스크립트로 GitHub PR 상태와 리뷰 대화, CI 실패를 확인합니다. PR 요약과 미해결 논의, 실패한 작업이나 단계의 로그 발췌를 살펴보고 전체 로그를 파일로 저장할 수 있습니다.

  • PR 상태를 한눈에 요약할 때
  • 미해결 리뷰 대화를 확인할 때
  • 실패한 CI 작업과 단계를 조사할 때
humanlayer/humanlayer1.3만2026년 10월 29일 업데이트

bump-version

무료요약 있음

지원 패키지의 릴리스 기준 버전을 올리고 관련 파일의 버전 표기를 동기화한 뒤, 생성된 버전을 확인합니다. 변경 사항의 커밋과 푸시, 병합 후 릴리스 절차도 다룹니다.

  • 릴리스용 패키지 기준 버전 올리기
  • 관련 파일의 버전 표기 동기화
  • 푸시 전에 생성 버전 확인
humanlayer/humanlayer1.3만2026년 10월 29일 업데이트

calldiff

무료요약 있음

Git 커밋, 브랜치 또는 작업 중인 코드의 함수 호출 구조를 비교합니다. 호출 트리를 표시하고 시작 함수에서 대상 함수까지의 경로를 추적하며, 22개 언어를 지원합니다.

  • 코드 변경 전후의 호출 관계 비교
  • 브랜치 간 호출 구조 검토
  • 시작 함수부터 대상 함수까지 경로 추적
humanlayer/humanlayer1.3만2026년 10월 29일 업데이트

chrome-devtools-cli

무료요약 있음

`chrome-devtools-mcp` CLI로 터미널에서 브라우저 작업을 자동화합니다. 페이지 요소를 확인하고 클릭하거나 양식을 입력하고 탭을 이동하며, 스크린샷을 저장하고 네트워크 요청, 콘솔 메시지, 성능, 메모리를 조사하도록 셸 명령 작성을 지원합니다.

  • 양식 입력과 페이지 이동 자동화
  • 페이지 또는 요소 화면 캡처
  • 네트워크 요청과 콘솔 오류 조사
humanlayer/humanlayer1.3만2026년 10월 29일 업데이트

humanlayer의 스킬 모두 보기

이 스킬의 문제 신고