marketing-brief
Use when the user wants to generate a marketing brief from a GitHub PR, git ref comparison, codebase feature, or freeform description of a feature or product change
AlemTuzlak · 已有 AI 审核记录
安装与使用
运行命令前先阅读来源与依赖说明。复制命令不会自动安装或执行。
npx skills add AlemTuzlak/skills --skill marketing-brief技能说明
name: marketing-brief description: Use when the user wants to generate a marketing brief from a GitHub PR, git ref comparison, codebase feature, or freeform description of a feature or product change
Marketing Brief Generator
Generate structured marketing briefs from code changes. Designed for non-marketers who need professional-quality briefs.
Input Resolution
Resolve the argument (if provided) in this order:
- Matches GitHub URL or
#\d+pattern → PR - Contains
...or..→ git ref range - Resolves to existing file/directory → codebase feature
- Otherwise → freeform text
If no argument is provided, ask: "What should I analyze? You can provide a PR URL/number, a git ref range (e.g. v1.0...v2.0), a file/directory path, or just describe the feature."
If multiple interpretations match, confirm with the user.
Process Flow
digraph marketing_brief {
rankdir=TB;
"Resolve input" [shape=box];
"Phase 1: Discovery" [shape=box];
"Understanding confirmed?" [shape=diamond];
"Sensitive items?" [shape=diamond];
"Get framing approval" [shape=box];
"Phase 2: Configure" [shape=box];
"Phase 3: Generate" [shape=box];
"Phase 4: Output" [shape=box];
"Resolve input" -> "Phase 1: Discovery";
"Phase 1: Discovery" -> "Understanding confirmed?";
"Understanding confirmed?" -> "Phase 1: Discovery" [label="no, revise"];
"Understanding confirmed?" -> "Sensitive items?" [label="yes"];
"Sensitive items?" -> "Get framing approval" [label="yes"];
"Sensitive items?" -> "Phase 2: Configure" [label="no"];
"Get framing approval" -> "Phase 2: Configure";
"Phase 2: Configure" -> "Phase 3: Generate";
"Phase 3: Generate" -> "Phase 4: Output";
}
Do NOT skip phases. Ask questions at a natural pace. Don't overwhelm, but don't artificially slow things down either. If the user answers multiple questions at once, accept their bundled answers and skip ahead.
If the user says "just pick defaults", "you choose", or similar, pick reasonable defaults based on context, state what you chose, and ask for a single confirmation before proceeding.
Never make assumptions without confirming with the user. Be conversational and guide them through the process.
Phase 1: Discovery
Step 1 — Analyze the input
| Input type | What to read |
|---|---|
| PR | Diff, PR description, review comments, commit messages (gh pr view, gh pr diff). For large PRs (20+ files), focus on user-facing changes. |
| Git refs | git diff and git log between the two refs. For large ranges, prioritize commit messages and user-facing changes. |
| Codebase feature | Read the specified files/directories. |
| Freeform text | Parse the user's description. If it lacks specifics (no feature name, no value prop, no context), ask the user to provide more detail or point to a specific file/PR. Fall back to open-ended questions only if they can't. |
User-facing changes include: new features, UI changes, API changes, performance improvements, bug fixes, and documentation updates. Internal changes include: refactors, test additions, CI changes, and dependency bumps. When uncertain, list what you found and ask the user which are relevant.
Error handling:
ghnot installed/authenticated → inform user, suggestgh auth login, offer alternative input type- Invalid PR number or git ref → tell user it wasn't found, ask to verify
- File/directory not found → tell user, ask for correct path
Step 2 — Read broader product context
Read these if they exist: README, docs/, package.json (or equivalent), any marketing/landing page references.
Goal: understand what the product is, who it's for, what it does.
If none of these exist, ask the user: "I couldn't find product context in the repo. Can you give me a brief description of the product — what it is and who it's for?"
Step 3 — Present understanding
Present a structured summary:
"Here's what I understand was built:"
- Feature A — short description
- Feature B — short description
- Feature C — short description
"Please confirm, correct, add, or remove items. Which of these should be included in the brief?"
Ask clarifying questions for anything ambiguous. Do NOT proceed until the user confirms the feature scope.
Step 4 — Flag sensitive items
Scan for: breaking changes, deprecations, security fixes, migration requirements, removed features, controversial decisions.
If found, flag each one:
"I detected some items that need careful framing:"
- Breaking change:
oldMethod()was removed- Deprecation: v1 API endpoints marked deprecated
"How would you like each of these framed in the brief?"
Do NOT proceed to Phase 2 until framing is agreed upon for every sensitive item.
Phase 2: Brief Configuration
Ask these questions:
Q1 — Audience: "Who will read this brief?" (you personally, marketing team, designer, stakeholders, other)
Q2 — Detail level: "How detailed should the brief be?" (concise/scannable, moderate, or detailed/prose). The detail level influences depth and tone, not hard word limits. Concise means short punchy sections. Detailed means full prose with depth.
Q3 — Competitors: "Are there any competitors you want me to position against? I'll also do my own web research."
Q4 — Conditional sections: Evaluate which optional sections are relevant based on the feature context. Present all relevant proposals in a single message, and let the user approve or decline each:
"I think [section] would be useful here because [reason]. Want me to include it?"
Optional sections and when to propose them:
| Section | Propose when |
|---|---|
| Tone & Voice Guidance | Audience is marketing team or designer who will produce copy |
| Visual/Asset Suggestions | Feature has UI component or is visually demonstrable |
| Timeline/Launch Window | Change is part of a larger release or has coordination needs |