agentleFS
lucasmaher-hashlucasmaher-hash/claude-skills/case-study [Plan, write and build portfolio design case studies]/SKILL.md

case-study

Skill0 starsChanged 4 months agoMITView on GitHub

Plan, write, critique and build portfolio design case studies (UX/UI, product, industrial design, apps, concepts). Use whenever the user wants to turn a project into a case study, rewrite or review a project page, plan the research/evidence for one, decide its structure, visuals, captions or charts, or asks for a good case-study reference/example. Triggers include "case study", "Fallstudie", "project page", "portfolio page", "how should I present this project", "is this case study good", "good case study reference".

19 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use audit.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • OpenAISkill

    audit

    Audit or critique a product flow, journey, workflow, funnel, onboarding path, checkout path, settings path, screen, or multi-step product experience by capturing screenshots first, then reporting UX, design, and accessibility findings inline from that evidence. Use Figma only when the user explicitly asks for a board. Use when the user asks to audit, review, critique, inspect, assess, analyze, evaluate, or give feedback on a product experience.

    Also in 3 projects that use this

The file

We split this skill into the rules it gives an agent

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

  • 0

    Common

    Most of the 18 edited versions we read keep these.

  • 1

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Interview first, never guess: Ask (one at a time, recommend an answer each time) what you cannot read from the project files: context and …

    line 20
  2. Very specific:

    If working for Lucas Maher: the German master notes live in his Obsidian vault (04 Ressourcen/Design & Craft/Case Study Aufbau.md, Design …

    line 86

“Common” is measured: kept by at least 6 in 10 of the 18 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

Case study

A case study is not a gallery of final screens. Reviewers are checking whether the designer can spot a problem, make defensible decisions and move a product forward. Process and reasoning count for more than polish. Everything in this skill serves that one test.

Hard facts to design around:

  • A portfolio gets 30–60 s on first pass; a single case study gets 5–10 min. Everything essential must survive the 30-second version (headings + captions + images alone).
  • ~800 words where every sentence counts beat 4,000. Target 900–1,100 words of prose.
  • A portfolio is judged by its weakest case study. Three to five strong ones, not eight.

Workflow

  1. Interview first, never guess. Ask (one at a time, recommend an answer each time) what you cannot read from the project files: context and dates, the user's exact role and each teammate's part, status (shipped / prototype / concept), the problem in their words, the biggest decision or pivot and why, what evidence exists (numbers, feedback, recordings, sketches, v1 screens), what must be left out, and voice. Record answers in a table.
  2. Find the spine. One sentence the whole page serves: "X lets [user] [do thing] — so [pain] stops." Separate the product insight (spine) from the best story (usually the pivot, which becomes the centrepiece section, not the spine).
  3. Choose the structure. See references/structure.md. Default skeleton: hero + one-liner + overview block → problem → the one idea → the key decision/pivot → system & craft (2–3 hard details) → result & learnings. Give each section exactly one job.
  4. Plan the evidence. For every section: what proves it? Use references/evidence.md. Each study gets why / how / measure / show / example shape / what it changes on the page. A study that changes nothing on the page does not get space.
  5. Plan the visuals. references/visuals.md: 30-second path, before/after, motion only for behaviour, one message per chart, captions that state findings.
  6. Write. references/writing.md: finding → decision, first person for decisions, say what did NOT change and why, at most three learnings.
  7. Critique before shipping with the checklist below. When reviewing someone's page, report the verdict first, then the 3–5 changes that matter most, then the rest.

For a full plan document, copy assets/plan-template.md and fill it in.

Non-negotiables

  • No invented research, numbers, personas or quotes. Example values in a plan are labelled as examples and replaced before shipping. n=1 stated plainly beats a faked study, and reviewers spot the difference. Unsourced statistics presented like outcomes are the most common flaw even in good portfolios (see the Calmsphere notes in references/exemplars.md).
  • Don't open with the aesthetic. Problem first, unless one sentence of problem follows the hook immediately.
  • Don't bury the key decision in a "challenges" section at the bottom. It gets the most space and the biggest image.
  • Name roles precisely in team work. Never claim the whole thing.
  • Don't explain the tech. The reader is hiring a designer: one caption covers the tools.
  • Name honest weaknesses (n=1, no metrics, AI-assisted build, never tested in context). Said first, they protect the rest of the page.

Checklist

  • The main point fits in one sentence, in the hero
  • Overview block answers: what is it · user problem · my role · what I designed · outcome/validation
  • Role exact; teammates' parts named
  • Fact bar of 2–4 checkable numbers or facts (no metrics → verifiable facts: "On the App Store", "3 iterations", "Solo, 6 weeks")
  • Problem is the strongest paragraph, source named honestly
  • Research condensed into one model that forces a decision (not a persona, not a decorative 2×2)
  • Key decision shown before/after, same crop and size
  • At least one dead end or setback shown
  • Every animation shows behaviour, not decoration
  • Every caption carries a finding
  • 30-second test passes (images + captions + headings tell the story)
  • Every number has a source and a timeframe; no outcome claimed that wasn't measured
  • At most three learnings, honestly phrased; open questions allowed
  • Under ~1,100 words of prose; no two paragraphs without an image

References

  • references/structure.md: canonical order, page skeleton, the overview block, section jobs
  • references/evidence.md: study template, cheap honest studies, projects without metrics
  • references/visuals.md: 30-second path, charts, before/after, motion, captions, rhythm, web specs
  • references/writing.md: voice, captions, sentence patterns, what to cut
  • references/exemplars.md: Calmsphere (go-to good example) + Jason Yuan ×2, Jessica Im, Deji Ajetomobi ×2, and lessons from real rebuilds
  • assets/plan-template.md: fill-in plan for one project

Lucas-specific context

If working for Lucas Maher: the German master notes live in his Obsidian vault (04 Ressourcen/Design & Craft/Case Study Aufbau.md, Design References tagged case-study-vorbild). Per-project plans live in ~/TEMP/3D_Portfolio/*casestudy*plan.md. His voice rules: 00 Kontext/Schreibstil.md. When he asks for a good case-study reference, start with Calmsphere (karenhappuch.com/calmsphere).

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.