본문으로 건너뛰기
cccskills
무료GitHub에 공개

product-design

Use when deciding what an interface should do before UI is built or audited: consequences, action scope, reachable states, naming, and accessibility. Not for visual implementation: use prototype.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

설치 방법 보기

포함된 파일(7)

  • SKILL.md6.3 KB
  • agents/openai.yaml217 B
  • references/interface-quality.md1.3 KB
  • references/naming-and-copy.md1.3 KB
  • references/product-judgment.md1.6 KB
  • references/rules.md2.7 KB
  • references/surfaces.md2.2 KB

SKILL.md(원문)

설치하기 전에 에이전트에게 전달되는 지시 내용을 확인할 수 있습니다.

Product design

Contract

FieldBound contract
Triggerproduct design, what should this do, interface should do, before anyone builds it, product requirements
AuthorityRead-only. No file, VCS, credential, paid, published, deployed, or remote mutation. Analyzes the user request, existing project files, design system artifacts, and AGENTS.md.
Side effectChat output only: product decisions, surface definitions, findings, and routed follow-on work.
DoneA report is returned defining what the interface does, its surfaces and reachable states, scope and consequence per action, findings with rule IDs, and routed follow-on work; or the report is labeled INCOMPLETE with the blocking gap named.

Inputs

Required:

  • The user's request text.

Required when available:

  • The project design system, existing UI, or AGENTS.md.
  • The surface or component under design.
  • Any existing spec, brief, or mockup.

Procedure

  1. Classify the request into one mode. Pick exactly one from the mode table below; when the verb is ambiguous, shape is the default. Modes chain: shape leads into spec; review leads into harden. Done when: one mode is named.

  2. Locate authority in this order: (a) the user's explicit goal and constraints, (b) verified user and product evidence, (c) project-canonical guidance in AGENTS.md and the design system, (d) this skill's standards. Load references/rules.md plus the mode's other reference files as listed in the mode table; rules.md loads in every mode. Every rule ID cited later must appear in a loaded file. Done when: the authority order is resolved and the mode's references are loaded.

  3. Write the internal brief (shape, spec, harden) before proposing UI: job, desired outcome, consequence, object, action scope, permissions, as defined in rules.md. If job, desired outcome, or consequence cannot be filled, stop and ask; do not propose UI. For spec and action, name the object, scope, and consequence for each action in scope. For shape, spec, and harden, enumerate every reachable state per surface against the checklist in surfaces.md and mark each reachable or unreachable with a reason. Done when: the brief is complete, per-action scope and consequence are named where required, and the reachable-state checklist is walked where required.

  4. Name scope and consequences and apply the loaded standards. Cite a rule ID for every finding and every non-mechanical decision. If no loaded rule governs a decision, record a coverage gap inline (three parts: proposed slug, the decision it would govern, category); never cite an invented ID. For review and harden, order findings P0-P3 by user impact; each finding carries location, verification status, rule ID, user consequence, the smallest concrete fix, and the accountable domain owner. Done when: every finding and non-mechanical decision carries a rule ID or a coverage gap, and review/harden findings are P0-P3 ordered with all six fields.

  5. Report findings and route follow-on work. Classify follow-on work by accountable domain: build changes belong to the implementation owner; exact wording to the copy and UX owner; passage timing to the motion owner; deep type treatment to the typography owner. Return that ownership map as data; do not invoke another skill. Then run the pass self-check: label the report INCOMPLETE if a finding lacks a rule ID, a cited ID was invented, or the internal brief is missing job, desired outcome, or consequence. Done when: the ownership map is returned and the self-check has run.

Modes

ModeDispatch when the user asks forLoad
shape (default)design the flow, what control here, how should this work, a brief with no settled UIrules.md, product-judgment.md
specspec the right interaction, define the expected statesrules.md, surfaces.md, naming-and-copy.md, product-judgment.md; mark implementation as a separate accountable domain
reviewreview this flow for product correctness, is this the right interactionrules.md, interface-quality.md
actionwhat should this action affect, which object or scope, action reversibility unsettledrules.md, naming-and-copy.md; classify wording polish under copy and UX ownership
hardenmake this resilient, what breaks hererules.md, surfaces.md, interface-quality.md, product-judgment.md

All paths are relative to this skill's references/ directory.

review mode judges the product decision, not rendered-artifact quality. For a component or UI audit, return a complete surface-audit brief with the target, primary task, reachable states, and verified product constraints; do not perform an unrelated visual implementation review here.

Failure and recovery

FailureResponse
Cannot fill job, desired outcome, or consequence in the internal briefStop and ask. Do not propose UI.
A rule ID is invented (does not appear in the loaded references)Record a coverage gap instead. Never cite a made-up ID.
Scope is ambiguousAsk. Do not widen scope without explicit user authorization.
Request spans multiple authoritiesReturn an ownership map by accountable domain. Finish this skill's product-decision scope and do not invoke another skill.

Partial-result rule: emit what is complete and label the report INCOMPLETE. Do not fabricate findings to fill the template.

Output

Product design report containing: the mode applied; the internal brief (shape, spec, harden): job, desired outcome, consequence, object, action scope, permissions; object, scope, and consequence for each action in scope (spec, action); the reachable-state checklist with each state marked reachable or not for this surface (shape, spec, harden); findings ordered P0-P3 by user impact, each with location, verification status, rule ID, user consequence, smallest concrete fix, and accountable domain owner (review, harden); coverage gaps recorded inline with proposed slug, decision, and category; the ownership map routing follow-on work to accountable domains; and the INCOMPLETE label when any self-check fails. Output length follows the work, not the template. Drop sections a pass did not need.

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

Use when explaining System V AMD64, ARM AAPCS, RISC-V psABI, stack frames, variadic calls, or FFI register rules. Not for the Rust FFI binding layer: use rust-ffi.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

OutlineDriven/odin-claude-plugin372026년 9월 29일 업데이트

Use when configuring ADC sampling time, DMA-driven ADC, calibration, or DAC channel setup on bare-metal MCUs. Not for the DMA stream itself: use dma-baremetal.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

OutlineDriven/odin-claude-plugin372026년 9월 29일 업데이트

af-xdp

무료

Use when creating AF_XDP sockets, configuring UMEM and XSK rings, writing an XDP redirect program, or choosing copy versus zero-copy mode. Not for full kernel bypass: use dpdk.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

OutlineDriven/odin-claude-plugin372026년 9월 29일 업데이트

Use when a completed session needs an agent-environment retrospective. Not for an engineering retrospective from telemetry: use engineering-retrospective.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

OutlineDriven/odin-claude-plugin372026년 9월 29일 업데이트

Use when a redacted, trimmed agent transcript must be appended to a GitHub PR or issue body, with human approval and preview. Not for automated or model-initiated insertion.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

OutlineDriven/odin-claude-plugin372026년 9월 29일 업데이트

agents-md

무료

Use when a repo needs agent setup, AGENTS.md added or made lean, CLAUDE.md audited, or agent instructions scored or pruned. Not for remote, credential, publish, deploy, or irreversible changes.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

OutlineDriven/odin-claude-plugin372026년 9월 29일 업데이트

OutlineDriven의 스킬 모두 보기

이 스킬의 문제 신고