Token导航 LogoToken导航TokenDH.com
研究检索权限需确认github未标认证来源可访问许可证需确认审计通过

briefingsbriefings 搜索

Agent Skill

briefings 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

总安装

706

周安装

30

GitHub Stars

55

下载量

247
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

unknown

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:briefings(briefings 搜索)
来源仓库:https://github.com/adobe/skills
仓库路径:skills/briefings
安装命令:
npx skills add https://github.com/adobe/skills --skill briefings
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/adobe/skills --skill briefings

简介

briefings 帮助表达页面愿景与业务意图,支持轻量句子到详细图文方向。

  • 适用于页面规划、受众定位与 CTA 映射。
  • 仅处理 stardust/briefings//*.md 文件变更。
  • 不涉及品牌身份或视觉规范,需配合其他工具使用。
  • briefings 属于研究检索类 Skill,可作为该场景下的辅助能力补充。

SKILL.md

Briefings

Help the user express their page vision as briefings that capture business intent. A briefing can be as lightweight as a single sentence of intent, or as detailed as final copy and image direction — the user chooses the fidelity.

When to use this skill

  • The user wants to plan pages, write briefings, define audience or CTAs, or map shared sections across pages.
  • The user asks to change, refine, review, critique, or iterate on any file under stardust/briefings/**/*.md (including _site.md).
  • The user types /stardust:briefings.

Do NOT use this skill

  • For brand identity (voice, colors, typography). Hand off to brand.
  • For structural wireframes or layout decisions. Hand off to wireframes or prototype.
  • To render pages. Hand off to wireframes (grey) or prototype (branded).

Pre-flight

Run the procedure in ../_shared/preflight.md first.

Contract

Needs (reads if present):

  • User description of pages to plan
  • .impeccable.md (tone hints)
  • stardust/brand-profile.json (voice; not required)

Produces:

  • stardust/briefings/_site.md (multi-page only)
  • stardust/briefings/{page}.md (one per page; sole source of truth for page copy)

If missing:

  • No input at all → ask the user which pages they need and start with the most important one.
  • No brand-profile.json → briefings still proceed; tone defaults to neutral-technical.

Copy Ownership

Briefings own page copy. If the user wants final words baked in, put them in # Copy per section. Design will use those strings verbatim and will never rewrite them. Briefings without # Copy are valid — downstream skills synthesize on-brand copy and stamp provenance.


Phase 1: Scope

If no briefings exist yet:

  1. Ask the user: "What pages do you need? Let's start with the most important one."
  2. For multi-page sites, create stardust/briefings/_site.md first — see briefing-format.md for the site briefing schema.
  3. Default to the structured shape for every page. Do NOT ask the user to pick a fidelity level. If a section's content is not known yet, emit [TBD] in that section — downstream skills synthesize on-brand content for [TBD] fields and stamp provenance. Fidelity is a consequence of how much the user fills in, not a gate the user has to cross. Deepening the fidelity ladder:

- Structured (default) — full frontmatter + Intent + Audience + Key Messages + CTAs + Tone. Any field may be [TBD]. - Fully specified — adds # Copy (final headlines/body/microcopy) and # Imagery (image direction per section, source hints, alt-text). Reached by filling in those sections, not by re-answering a fidelity question.

If briefings already exist, read stardust/briefings/ and confirm which pages are covered. Ask which page to work on next (new, edit, or deepen by filling in [TBD] fields or adding # Copy / # Imagery).

Phase 2: Draft

  1. Run the soft-deps discovery decision before any interview question: If /brainstorm is registered in this session (detect per ../_shared/soft-deps.md): hand off to /brainstorm with a seeded prompt that includes the page name and any existing briefing content, and ask it to produce the structured shape (Intent + Audience + Key Messages + CTAs + Tone). Wait for its output; use the result to populate the briefing. Do NOT run any inline interview question once this delegation is made — that would double-interview the user. Otherwise (superpowers not installed): announce the fallback exactly once per session, using the verbatim text from ../_shared/soft-deps.md ("superpowers announcement"). Then run the "For briefings discovery" pattern in ../_shared/fallback-brainstorm.md. On every subsequent briefing in the same session, skip the announcement (already seen) and run the inline pattern directly. Either path uses the same discovery prompts: "What should visitors feel when they land on this page? What's the one action you want them to take?" The user must be able to tell which path ran — either by /brainstorm's visible hand-off UI or by the one-time announcement. Silent inline interviews are a bug.
  2. Draft the briefing in the structured shape. Fill every section; where the user has not provided content, write [TBD] verbatim (not a rephrased placeholder). The agent helps draft; the user owns the content.
  3. Present the draft and wait for approval before writing the file.
  4. Write approved briefings to stardust/briefings/{page}.md.

If any [TBD] fields remain, or the briefing was synthesized from a short prompt rather than a full conversation, stamp a provenance comment at the top of the file per ../_shared/skill-contract.md so downstream skills know which fields to treat as synthesizable.

Phase 2.5: Content Reuse Map (Multi-Page Sites Only)

Before the user moves on to wireframes, plan how pages connect through shared content. This is not just cross-links — it's content reuse: the same recipe card, testimonial quote, or capability highlight appears on multiple pages, authored once.

  1. Identify the main content types each page owns (e.g., recipes page owns recipe cards, stories page owns testimonials).
  2. Map where each content type gets reused as excerpts on other pages.
  3. Add a # Content Reuse Map section to stardust/briefings/_site.md — see briefing-format.md for the table format.
  4. Present the map to the user for approval before closing the stage.

Key principles:

  • The homepage is a hub — it should pull excerpts from every major content page.
  • Inner pages cross-link to at least 2 siblings through reused content sections (not just nav links).
  • Reused sections show 3–4 items from a source that has more, creating "see more" motivation.
  • Every reused section includes a CTA to the source page.

Phase 3: Approval Gate

This is a soft gate. The user may want to write briefings for some pages now and others later.

  • After each briefing is approved, ask: "Another page, or are we done for now?"
  • When the user is done, tell them:

- If stardust/brand-profile.json exists: "Briefings saved. You have brand + briefings — you can now run /stardust:wireframes." - If not: "Briefings saved. Run /stardust:brand when you're ready to capture your brand, then /stardust:wireframes to render these pages."

Pipeline automation: When invoked as part of a full pipeline run, auto-approve each briefing and continue.

Why Briefings Are Standalone

Briefings capture *what the experience should be* — intent, audience, messages. That work does not depend on having a brand. A user can:

  • Write briefings first, then extract brand, then wireframe.
  • Extract brand first, then write briefings, then wireframe.
  • Write briefings without ever running brand (useful for scoping conversations or paper-only planning).

Keeping briefings independent lets teams split the work: one person captures intent while another extracts the brand.

Artifacts Written

FileDescription
stardust/briefings/_site.mdSite-level briefing (multi-page only)
stardust/briefings/{page}.mdPage-level briefings in the structured shape (Intent + Audience + Key Messages + CTAs + Tone), with [TBD] placeholders where content is not yet captured. May additionally contain # Copy / # Imagery when the user has committed to final words or image direction.

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Codex

35.05%
按下载量换算87

Claude

28.14%
按下载量换算70

Cursor

19.05%
按下载量换算47

Gemini CLI

9.53%
按下载量换算24

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

权限需确认

当前来源未能明确判断权限范围,默认进入异常复核队列。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills