Purpose
Translate a messy, one-line, or partial creative request into a complete, buildable design specification — surfacing the objective, audience, format constraints, messaging priority, and visual direction — so a designer can start work with confidence instead of chasing clarifications for two days.
When to Use
Invoke this skill when the user asks to:
- Turn a vague marketing or sales request into a real design brief
- Figure out what a stakeholder actually wants before designing
- Fill in the gaps of an incomplete or one-sentence brief
- Structure a brief that arrived as a Slack message, email, or verbal ask
- Identify what information is missing before starting a design
- Standardize how inbound design requests get documented
Inputs Required
Before starting, collect:
- Raw brief: The request as received — paste it verbatim, however incomplete
- Deliverable type: What's being made (social post, landing page, one-pager, ad set, deck, banner, etc.)
- Requesting stakeholder: Which function it came from (marketing, sales, exec, product)
- Deadline: When it's needed, or "unknown"
- Existing brand assets: Whether a brand guide, logo files, or prior work exist
- Known audience: Who it's for, if stated
- Distribution channel: Where it will appear, if stated
If deliverable type, audience, or channel is missing, infer the most likely option from the raw brief and flag it as an assumption to confirm — never invent a fixed constraint silently.
Instructions
Step 1: Extract What's Actually Stated
Read the raw brief and separate explicit facts from implied ones. Create two short lists:
- Stated: Direct quotes or clear intent from the brief
- Implied: Reasonable inferences (e.g., "for the sales team's outreach" implies a B2B decision-maker audience) Do not treat implications as confirmed. This separation drives the Open Questions section later.
Step 2: Define the Objective
Reframe the request as a single measurable objective. Push past "make a graphic" to the business outcome:
- Action objective: What should the viewer do? (click, sign up, remember, feel, share)
- Success signal: How the stakeholder would know it worked Write one sentence: "This piece should get [audience] to [action] so that [business outcome]."
Step 3: Profile the Audience
Convert vague audience language into a usable design profile:
- Who: Role, segment, or persona
- Awareness stage: Cold / warming / ready-to-buy — affects how much explanation the design must carry
- Reading context: Mobile scroll, desk research, in a meeting, print — affects density and hierarchy
- What they care about: The one thing that earns their attention in this format
Step 4: Lock the Format Specs
Turn the deliverable type into hard, buildable specs. Provide concrete defaults for the channel and flag any that need confirmation:
- Dimensions: Exact px or print size (e.g., 1080×1350 for IG portrait, 1200×628 for LinkedIn link ad)
- Format/file type: PNG, MP4, PDF, HTML, etc.
- Orientation and safe zones: Where text and logos must sit
- Length constraints: Character limits, aspect ratios, animation duration
- Variants needed: How many sizes/versions (e.g., story + feed + banner)
Step 5: Build the Messaging Hierarchy
Rank the content so the designer knows what to make loud and what to make quiet:
- Primary message (hero): The one thing that must land in under 3 seconds
- Secondary support: Proof, benefit, or context (one tier down)
- Tertiary / functional: CTA, disclaimers, logo, URL Note the exact copy if provided; if copy is missing, specify what each tier needs to communicate.
Step 6: Set the Visual Direction
Give enough direction to start without over-prescribing:
- Tone: 3–4 adjectives (e.g., confident, clean, human)
- Reference points: Named styles or comparable work if implied by the brief
- Brand constraints: Colors, type, logo usage — pull from existing assets if available, otherwise flag
- Do / Don't: 2–3 specific guardrails (e.g., "Do use real product screenshots" / "Don't use stock handshake photos")
Step 7: List Deliverables and Timeline
- Enumerate every file the stakeholder will actually receive
- Note round expectations (concepts → revision → final)
- Restate the deadline or flag it as unresolved
Step 8: Surface Open Questions
List every gap that would block or misdirect the work — drawn directly from the "Implied" list in Step 1. Frame each as a specific, answerable question, not a vague concern. Rank by how much the answer changes the design.
Output Format
- Structure: A single structured brief document with labeled sections, scannable
- Length: 500–800 words
- Format: Structured document — short labeled sections with bullets, not prose
- Sections: Objective → Audience Profile → Format Specs → Messaging Hierarchy → Visual Direction → Deliverables & Timeline → Open Questions to Resolve → Assumptions Made
- Tone: Design lead reviewing intake — precise, decisive, no filler; states defaults confidently and flags gaps plainly
Quality Checklist
Before delivering output, verify:
- [ ] The objective is written as a single "get [audience] to [action] so that [outcome]" sentence
- [ ] Format specs include exact dimensions and file type — not "standard social size"
- [ ] Messaging hierarchy has exactly three ranked tiers (hero / secondary / tertiary)
- [ ] Every assumption made is listed explicitly in the Assumptions section
- [ ] Open Questions are phrased as specific answerable questions, ranked by design impact
- [ ] Visual direction includes at least 2 concrete Do/Don't guardrails, not generic adjectives alone
Notes
- The most common failure in inbound briefs is a missing objective — stakeholders describe the artifact, not the outcome. Always reframe to outcome first.
- When the raw brief is a single sentence, the Open Questions section is the most valuable output — it converts a follow-up meeting into a checklist.
- Provide real dimension defaults per channel rather than asking; a designer can start on the likely spec and adjust, which is faster than blocking on confirmation.
- Keep assumptions and open questions strictly separate: assumptions are what you proceeded with; open questions are what could still redirect the work.
How to use this skill
This file is a Claude Code skill in Anthropic's standard format — markdown with YAML frontmatter at the top.
Installing in Claude Code:
~/.claude/skills/<folder-name>/SKILL.md— auto-loads in every session anywhere on your machine (recommended for skills you'll use across multiple projects)<your-project>/.claude/skills/<folder-name>/SKILL.md— auto-loads only when working in that project (recommended for project-specific skills)
Restart Claude Code or open a new session to register.
Using in Claude.ai, ChatGPT, Gemini, or other AI tools:
- Drag this file into a new chat, or copy and paste its contents
- Send your task or request as your next message — the AI will apply the skill where it fits
For Claude Code skill details: https://docs.anthropic.com/en/docs/claude-code/skills