case-study
Skill1 starsChanged 6 months agoApache-2.0View on GitHub
Generate a case study page that documents a specific client outcome — problem, intervention, measurable result, and the testable claim. Captures decision-stage searchers looking for proof of capability ("[industry] [outcome] case study", "[service] results"). Pulls the real story from the site's per-site `stories.md` (refuses if no eligible story exists — never invents). Satisfies Tier 1 on-page SEO, validates schema, and ships per the site's publishing method. Use when the user types `/case-study` or asks for a case study / customer story page.
19 other projects use a version of this.
Adopted by
Most recently taken up by
- matteotitta/genesys-skillstheir own version, first seen Nov 15, 2026
- BeautifulEngine/case-studytheir own version, first seen Nov 15, 2026
- tidycatc132/claude-skillstheir own version, first seen Nov 8, 2026
- Jakeschincariol/whale-skilltheir own version, first seen Nov 8, 2026
- lucasmaher-hash/claude-skillstheir own version, first seen Nov 1, 2026
- MakiDevelop/claude-skillstheir own version, first seen Nov 1, 2026
Other versions19 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 19
- duncan-buildroom/freeskills★ 65same name
- darkroomengineering/cc-settings★ 53same name
- matteotitta/genesys-skills★ 53same name
- launchapp-dev/animus-cli★ 39same name
- ommakes/Skills★ 21same name
- SonwaneyY/portfolio-case-study-generator★ 17same name
- Jakeschincariol/whale-skill★ 15same name
- calven-ai/marketing-as-code★ 10same name
- scoobynko/case-study-skill★ 9same name
- kronus-tech/kronus★ 4same name
9 more not listed
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 repository20 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- alternativeSkill
- auditSkill
- blogSkill
- broken-backlinksSkill
- clusterSkill
- comparisonSkill
- competitor-backlinksSkill
- glossarySkill
- haroSkill
- integrationsSkill
- lovable-deploySkill
- pricingSkill
- programmatic-batchSkill
- refreshSkill
- semrush-baselineSkill
- serp-featuresSkill
- serviceSkill
- triageSkill
- wireframeSkill
The file
We split this skill into the rules it gives an agent
6 rules. 0 are common, 1 is distinctive, and 5 are 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.
5
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 27
Read references/stories.md (per-site, per-language).
- Distinctive: line 30
If any required field is missing → refuse, escalate or ask.
- Very specific: line 55
If the client is named with consent: Review with itemReviewed = our service, reviewBody = the quote, author = the client.
- Very specific: line 56
Do NOT use Review schema for anonymized cases — schema requires identifiable author.
- Very specific: line 59
Every number on the page MUST be in stats.md.
- Very specific: line 67
No matching story exists in stories.md for the given ID.
“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.
Files
/case-study — Client outcome case study page
Documents one real client outcome end-to-end. Case studies sit at the decision stage — readers are evaluating whether the site can produce the result they need. Every claim must be verifiable; every number must be sourced from stats.md; every quote must be from a real, consented testimonial. No invention — refusal is the correct response when there's no eligible story.
Inputs
Required (asked at start if not given):
- Site — must match a folder under
sites/[name]/. - Story ID — references the entry in
sites/[site]/references/stories.md(orstories.{lang}.mdfor multilingual). Format:story-{slug}matching the heading anchor in stories.md.
Optional (per-run overrides):
- Industry / vertical — used as the H1's qualifying detail and the schema
industryfield. Defaults to the story's industry. - Anonymize —
true(default) |false(only if the story file marks the client as having given explicit consent for name use). If consent flag is missing, anonymize is forced true.
Workflow
This skill follows /blog's workflow (Steps 0–11) with the page-type-specific differences listed below. Read .claude/skills/blog/SKILL.md end-to-end before running.
What's unique to /case-study
Step 2 — Story selection (not keyword selection):
- Read
references/stories.md(per-site, per-language). - The user supplies the story ID, or the skill picks the highest-impact unused story (one with the largest measurable outcome).
- Required fields in the story entry:
client_descriptor,problem,intervention,outcome_metric(must be a number fromstats.md),consent(anonymizedornamed-with-consent). - If any required field is missing → refuse, escalate or ask.
Step 3 — Keyword cluster (case-study-specific):
- The primary keyword is the outcome-search query, not the story:
[industry] [outcome] case study,[service] results [region],how [client-type] achieved [outcome]. - Pick from
keywords.csvrows with intent=case-studyorproof. If none exist, the skill suggests one based on the story's industry + outcome and asks the user to add it to keywords.csv before proceeding.
Step 4 — SERP analysis (case-study-specific):
- Top 3 case study pages in this niche — extract: average word count, image count, presence/absence of dollar figures, presence/absence of named clients, format (long-form vs. multi-section vs. PDF-download-gated).
- Most ranking case studies are NOT gated. The default is unaffected by gating decisions made elsewhere.
Step 6 — Wireframe (case-study-specific):
- Hero: the outcome stated upfront as a number. "[Client descriptor] achieved [outcome number] in [timeframe]" — no marketing fluff before the number.
- Quick fact strip (above fold): 3–4 data points —
Client: [descriptor],Industry: [vertical],Timeframe: [duration],Outcome: [metric]. Skimmable. - Problem section (300–500 words): what the client was experiencing. Specific. With dates if possible. Avoid generic "they wanted to grow."
- Intervention section (400–700 words): what we did, in chronological order. Specific tools, frameworks, hours invested if known. This is where AI-search-citation density matters most — readers + LLMs both pull from this section.
- Result section (300–500 words): the outcome metric, the time it took, what changed for the client beyond the headline number.
- Quote (required if
named-with-consent, optional if anonymized): one verbatim sentence from the client. Cited. - "What we'd do differently" section (required, 100–200 words): one thing about the engagement that didn't work or that we'd change. Single biggest trust-building section in the format.
- FAQ: 4–6 Qs — "Was this typical?", "How transferable is this?", "Pricing on engagements like this?"
- CTA: softer — "See if your situation maps to this one" rather than "Book a call."
Step 7 — Schema (case-study-specific):
Articlewithabout: { @type: "CreativeWork", name: "[Engagement name]" }andmentionsfor the industry.BreadcrumbListFAQPage- If the client is named with consent:
ReviewwithitemReviewed= our service,reviewBody= the quote,author= the client. - Do NOT use
Reviewschema for anonymized cases — schema requires identifiable author.
Step 8 — Voice (case-study-specific extras):
- Every number on the page MUST be in
stats.md. The skill refuses to fabricate or round. - No "incredible results" or "transformative outcome" — banned per universal voice rules. Use the number.
- The "what we'd do differently" section is non-skippable. If the story file doesn't have material for this, refuse and ask the user to add the introspection.
- One opinion max, backed by the outcome metric.
Refusal conditions
This skill refuses to ship if:
- No matching story exists in
stories.mdfor the given ID. - The story is missing
outcome_metric(must be a number, must be instats.md). - The story is missing
consentfield. - The "what we'd do differently" content is unwritable from the story file (no introspection available).
- A case study for this story already exists in
used-keywords.md.
Output
Standard _drafts/[slug]/ layout, with these additions:
story-source.md— the source story entry from stories.md, copied for traceability.numbers-citation.md— every number on the page mapped to itsstats.mdline.
Length target: 1,500–2,500 words.
Routine versioning + idempotency contract
Same as /blog. Idempotency key: (site, story-id, language). Skip re-run if _drafts/case-study-{slug}/ exists with today's date.
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
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.

