design-flow
Skill536 starsChanged 7 months agoApache-2.0View on GitHub
Run the full design-to-build workflow as a guided sequence. Orchestrates all designer skills in order, from grilling through review. Use when user wants to go through the complete design process, start a project from scratch, run the full flow, or mentions "design flow" or "full workflow".
We split this skill into the rules it gives an agent
9 rules. 0 are common, 1 is distinctive, and 8 are very specific to this project.
0
Common
Most of the 7 edited versions we read keep these.
1
Distinctive
This author's own choices. Worth a look.
8
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 43
The designer can stop at any point: If they say "that's enough for now," summarize where they are in the sequence and what the next phase …
- Very specific: line 49
Read the grill-me skill (grill-me/SKILL.md) and follow its instructions.
- Very specific: line 55
Read the design-brief skill (design-brief/SKILL.md) and follow its instructions.
- Very specific: line 61
Read the information-architecture skill (information-architecture/SKILL.md) and follow its instructions.
- Very specific: line 67
Read the design-tokens skill (design-tokens/SKILL.md) and follow its instructions.
- Very specific: line 73
Read the brief-to-tasks skill (brief-to-tasks/SKILL.md) and follow its instructions.
- Very specific: line 79
Read the frontend-design skill (frontend-design/SKILL.md) and follow its instructions.
- Very specific: line 95
When triggered, read the design-review skill (design-review/SKILL.md) and follow its instructions.
- Very specific: line 102
All design flow artifacts are saved under .design/<feature-slug>/ where <feature-slug> is a short, lowercase, hyphenated name derived from …
“Common” is measured: kept by at least 6 in 10 of the 7 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
This skill orchestrates the full designer workflow by running each skill in sequence. You are a guide walking the designer through each phase. Do not rush. Each phase must be completed and confirmed before moving to the next.
Example prompts
- "Run the full design flow"
- "Walk me through the complete process for a new project"
- "Start from scratch and take me through everything"
- "Design flow for a dashboard app"
The Sequence
1. Grill Me → Clarify thinking
2. Design Brief → Document intent
3. Info Architecture → Define structure
4. Design Tokens → Establish visual system
5. Brief to Tasks → Plan the build
6. Frontend Design → Build it
—
7. Design Review → Run separately when readyRules
- At the start, tell the designer what the full sequence looks like (phases 1-6, with review available separately) and ask if they want to skip any phases. Common skip patterns:
- Already have a clear idea → skip grill-me
- Single component, not a full page → skip information-architecture
- Existing project with tokens → skip design-tokens
- Before each phase, announce which phase you are entering and what it will produce. Example: "Phase 2: Design Brief. I'll interview you about the project and produce a DESIGN_BRIEF.md file. Ready?"
- During each phase, read the corresponding SKILL.md file and follow its full instructions. Do not summarize or abbreviate the skill. Run it properly.
- After each phase, summarize what was produced (the file name, the key decisions, any open questions) and ask: "Ready to move to the next phase?" Wait for confirmation.
- Between phases, check if the output from the previous phase changes anything about the next phase. For example, if the brief names a philosophy, mention that the tokens phase will use it.
- The designer can stop at any point. If they say "that's enough for now," summarize where they are in the sequence and what the next phase would be when they return.
Phase Details
Phase 1: Grill Me
Read the grill-me skill (grill-me/SKILL.md) and follow its instructions.
Produces: Shared understanding of the project. No file output.
Transition: "We've resolved the key decisions. Ready to capture this as a design brief?"
Phase 2: Design Brief
Read the design-brief skill (design-brief/SKILL.md) and follow its instructions.
Produces: .design/<feature-slug>/DESIGN_BRIEF.md.
Transition: "The brief is saved. Next is information architecture, where we'll define the page structure and navigation. Skip this if you're building a single component. Continue?"
Phase 3: Information Architecture
Read the information-architecture skill (information-architecture/SKILL.md) and follow its instructions.
Produces: .design/<feature-slug>/INFORMATION_ARCHITECTURE.md.
Transition: "IA is defined. Next we'll generate design tokens (colors, spacing, typography) based on the philosophy from the brief. Continue?"
Phase 4: Design Tokens
Read the design-tokens skill (design-tokens/SKILL.md) and follow its instructions.
Produces: Token file (CSS variables, Tailwind config, or theme file depending on stack).
Transition: "Tokens are set. Next I'll break the brief into a task list so we can build in order. Continue?"
Phase 5: Brief to Tasks
Read the brief-to-tasks skill (brief-to-tasks/SKILL.md) and follow its instructions.
Produces: .design/<feature-slug>/TASKS.md.
Transition: "Tasks are ready. Now we build. I'll start with the first task on the list. Continue?"
Phase 6: Frontend Design
Read the frontend-design skill (frontend-design/SKILL.md) and follow its instructions.
Work through the tasks from TASKS.md in order. After completing each task, check it off and confirm with the designer before moving to the next task.
Produces: Built components and pages.
Transition: "The flow is complete. Your brief, IA, tokens, and tasks are all saved in the project. When you're ready for a design review, run /design-review and I'll critique the build against the brief."
The flow ends here. Phase 7 is not automatic.
Phase 7: Design Review (on request only)
This phase does NOT run automatically. It only runs if:
- The designer explicitly asks for a review during the flow
- The designer runs
/design-reviewseparately after building
The review requires built code to examine. If no components or pages have been built yet, do not run this phase. Instead, remind the designer: "Run /design-review once you have something built. It will check the output against the brief."
When triggered, read the design-review skill (design-review/SKILL.md) and follow its instructions. The review will capture screenshots of the running application using Playwright MCP (preferred), the Cursor IDE Browser (fallback), or by asking the user to provide them manually if no browser tool is available.
Produces: .design/<feature-slug>/DESIGN_REVIEW.md + screenshots saved in .design/<feature-slug>/screenshots/.
Transition: "Review is done. Screenshots are saved in .design/<feature-slug>/screenshots/. If there are must-fix items, I can address them now."
Project Files Structure
All design flow artifacts are saved under .design/<feature-slug>/ where <feature-slug> is a short, lowercase, hyphenated name derived from the feature being designed. This ensures multiple features can be designed independently without overwriting each other.
.design/
└── <feature-slug>/
├── DESIGN_BRIEF.md ← Phase 2: Project intent, goals, aesthetic direction
├── INFORMATION_ARCHITECTURE.md ← Phase 3: Navigation, page structure, user flows
├── DESIGN_TOKENS.* ← Phase 4: Colors, spacing, typography, shadows (CSS/Tailwind/theme)
├── TASKS.md ← Phase 5: Ordered build checklist from the brief
├── DESIGN_REVIEW.md ← Phase 7: Prioritized critique against the brief
└── screenshots/ ← Phase 7: Visual evidence from the running app
├── review-[page]-desktop-1280.png
├── review-[page]-tablet-768.png
├── review-[page]-mobile-375.png
├── review-[page]-dark-mode-*.png
└── review-[component]-[state].pngThe screenshots/ subfolder is created during the design review phase. All visual evidence of the review (responsive breakpoints, interactive states, dark mode) is saved here with descriptive filenames so findings in DESIGN_REVIEW.md are traceable.
If the Designer Returns Mid-Flow
Check the .design/ folder for existing feature subfolders. If files from earlier phases exist (DESIGN_BRIEF.md, INFORMATION_ARCHITECTURE.md, TASKS.md) inside a feature folder, read them to understand where the designer left off. Ask which feature to resume if multiple folders exist. Resume from the next incomplete phase.
Other versions8 projects use a version of this.
Word for word · 1
Edited · 2
- drinkspiller/skills★ 1100% same
- hakushiidev/claude-skills★ 099% same
Wrote their own version · 5
- chipweaver/veripower★ 8same name
- viknesh20-20/claude-code-tool-kit★ 7same name
- DanielNoamTuby/claude-skills★ 0same name
- rynbol/claude-skills★ 0same name
- taro1037-design/claude-skills★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository7 other files it gives its agents.
- brief-to-tasksSkill
- design-briefSkill
- design-reviewSkill
- design-tokensSkill
- frontend-designSkill
- grill-meSkill
- information-architectureSkill
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
Reviews can't be read right now.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

