agentleFS
slogsdonslogsdon/skills-workflows/skills/workflow-design/SKILL.md

workflow-design

Skill0 starsChanged 3 months agoMITView on GitHub

Run the full design pipeline for a brand or one-off artifact — design-plan (strategy) → design-system (tokens + showcase) → an artifact skill (linkedin / blog-hero / etc.) → screenshot-html (PNG export). Use when the user says "design a brand", "I need the full design system + assets", "set up a new brand from scratch", "design pipeline for X", or when starting visual work on a brand that has no DESIGN.md yet. Do NOT use when the brand already exists and you just need one artifact — call the artifact skill directly.

6 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

4 rules. 0 are common, 1 is distinctive, and 3 are very specific to this project.

  • 0

    Common

    Most of the 5 edited versions we read keep these.

  • 1

    Distinctive

    This author's own choices. Worth a look.

  • 3

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Default: run plan → system → artifacts → screenshots in order, with checkpoints between each.

    line 60
  2. Very specific:

    If <step> is not in {plan, system, artifacts, screenshots}, print:

    line 64
  3. Distinctive:

    Before running the resume step, confirm the listed inputs exist.

    line 73
  4. Very specific:

    If any stage errors, stop and print: which step failed, the underlying skill's error, and the resume command (/workflow-design …

    line 97

“Common” is measured: kept by at least 6 in 10 of the 5 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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Skill: /workflow-design

Orchestrates the four-stage design pipeline. Replaces the "which design skill comes first" question with a single invocation.

When to use

  • Brand has no ./design/<slug>/DESIGN.md yet AND user wants one or more artifacts.
  • User explicitly asks for "the full pipeline" / "set up the brand from scratch".

When NOT to use

  • Brand already has a design system; just need one artifact → call the artifact skill directly.
  • User wants to iterate on existing tokens → call design-system directly.

Steps

Step nameStageInvokes
planStrategic intentdesign-plan
systemToken systemdesign-system
artifactsArtifact generationdesign-<artifact> (varies)
screenshotsPNG exportscreenshot-html

Stage 1: Strategic intent — step plan — invokes design-plan

Asks the brand-strategy questions (audience, voice, hard NOs, mood references) and writes ./design/<slug>/DESIGN-PLAN.md.

Output: ./design/<slug>/DESIGN-PLAN.md exists. Checkpoint: confirm DESIGN-PLAN.md before generating tokens.

Stage 2: Token system — step system — invokes design-system

Reads DESIGN-PLAN.md and produces DESIGN.md (Google design.md spec format), tokens.css, and showcase.html.

Output: three files in ./design/<slug>/. Checkpoint: confirm showcase.html renders correctly in a browser before generating any artifacts.

Stage 3: Artifact generation — step artifacts — invokes the artifact skill the user named

Asks: "Which artifact(s) do you want to generate? (linkedin-post / instagram-post / blog-hero / quote-card / business-card / podcast-cover / youtube-thumbnail / twitter-card / newsletter-header / talk-slide / speaker-bio-card / twitch-panels / obs-alert-overlay / obs-scene-pack / stream-overlay / link-in-bio / landing-page / ui-components / carousel-slide / youtube-channel-art / audit-report)"

For each chosen artifact, invokes the matching design-<artifact> skill with the brand slug. Produces an HTML file in ./design/<slug>/artifacts/.

Output: one or more .html files in artifacts/. Checkpoint: confirm artifacts before screenshotting.

Stage 4: PNG export — step screenshots — invokes screenshot-html

Runs the screenshot-html script against ./design/<slug>/artifacts with output to ./design/<slug>/screenshots.

Output: PNG files matching the artifact set.

Resume protocol

Default: run plan → system → artifacts → screenshots in order, with checkpoints between each.

--resume-from <step>: skip every step before <step> and begin from <step>. The named step itself runs.

If <step> is not in {plan, system, artifacts, screenshots}, print:

Invalid resume-from step: <step>
Valid steps: plan, system, artifacts, screenshots

…and stop.

Required state when resuming

Before running the resume step, confirm the listed inputs exist. If anything is missing, prompt the user for it (paste, file path, or re-run from an earlier step).

  • --resume-from plan — same as a fresh run; needs only the brand slug.
  • --resume-from system — needs ./design/<slug>/DESIGN-PLAN.md. If absent, ask the user to paste the plan or run from plan.
  • --resume-from artifacts — needs ./design/<slug>/DESIGN.md, ./design/<slug>/tokens.css, and ./design/<slug>/showcase.html. Ask the user which artifact(s) to generate before proceeding.
  • --resume-from screenshots — needs at least one .html file in ./design/<slug>/artifacts/. List the files found and confirm before screenshotting.

Examples:

/workflow-design --resume-from system     # already have a DESIGN-PLAN.md, generate tokens
/workflow-design --resume-from artifacts  # tokens look good, generate the LinkedIn post
/workflow-design --resume-from screenshots --no-pause  # batch the PNG export

Other flags

  • --skip <step> to skip a step (e.g. --skip screenshots if user only wants HTML).
  • --no-pause to run end-to-end without checkpoints.

Stage quality gate

After each stage: does the output meet the stage goal (plan = strategy clear; system = tokens render correctly; artifacts = visually match intent; screenshots = clean PNGs)? If it partially misses, retry the stage once with an explicit note on the gap. On second failure, surface to the user rather than auto-continuing.

Failure handling

If any stage errors, stop and print: which step failed, the underlying skill's error, and the resume command (/workflow-design --resume-from <step>).

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

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.