case-study

Generate evidence-backed engineering case studies from a repository's git history, measurements, and interviews.

Updated Mar 23, 2026
One-click install
npx skills add https://github.com/Haseeb-Arshad/my-folio-book --skill case-study-haseeb-arshad
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: case-study
Source: https://github.com/Haseeb-Arshad/my-folio-book/tree/main/.agents/skills/case-study
Command: npx skills add https://github.com/Haseeb-Arshad/my-folio-book --skill case-study-haseeb-arshad

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires @playwright/test, and includes references (resource) and assets (resource) components.

What problem does it solve? Writing a portfolio case study about employer or client work is risky and time-consuming: claims lack evidence, numbers lack methods, and screenshots can leak confidential data. This Skill runs a structured nine-phase workflow that turns a repository into a publishable, evidence-backed case study while enforcing confidentiality clearance at every step. ## Core Features & Use Cases - Evidence-first workflow: Mines git history, PRs, migrations, and code structure into a claim-by-claim evidence ledger before any prose is written, then interviews the user only for what the repo cannot answer. - Measurement and research phases: Benchmarks before-and-after states with git worktrees, Lighthouse, autocannon, and EXPLAIN ANALYZE, then anchors numbers against public baselines like Core Web Vitals thresholds. - Confidentiality gating and safe visuals: Assigns a disclosure tier, maintains a never-publish list, captures screenshots and short recordings against seeded synthetic data via a Playwright script, and runs a redaction and secret-scan QA pass. - Use Case: An engineer wants to publish a portfolio write-up of a dashboard performance project. The Skill clears what may be disclosed, proves the p95 improvement with measured benchmarks, captures redacted screenshots, drafts the case study, and packages a project.json and excerpt for the portfolio site. ## Quick Start Ask the assistant to write a portfolio case study for this repository using the case-study skill, starting with the clearance questions.

Frequently Asked Questions about case-study

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I write an engineering case study for my portfolio?▼

Run the case-study workflow from inside the project's repository. It mines git history for evidence, interviews you for context the repo lacks, measures real numbers, then drafts a 1,200 to 2,000 word case study with labeled metrics and captioned assets.

How do I write about employer work without violating an NDA?▼

The workflow starts with a clearance phase that assigns a disclosure tier: public, internal-safe, or restricted. It records a never-publish list, substitutes identifying names, uses seeded synthetic data for all screenshots, and recommends manager sign-off before publishing.

Can I use this if the project cannot run locally?▼

Yes, use the --fast mode, which skips measurement, research, and capture phases. The resulting case study must state plainly that its numbers are reported rather than measured, since no local benchmarks could be run.

How are performance numbers verified in the case study?▼

The measurement phase checks out the before-state with git worktree and benchmarks both sides on the same machine using tools like autocannon, Lighthouse, hyperfine, or EXPLAIN ANALYZE. Every number is labeled measured, reported, or estimated with method, environment, and date.

What prevents confidential data from leaking into screenshots?▼

All captures run against seeded synthetic data with a fictional demo user in a clean browser profile, using a Playwright script that freezes the clock and fails if a non-example email appears. A redaction pass and secret grep over the case-study folder run before anything is committed.