seed-skills/qa-project-portfolio-builder/SKILL.md
Turn test suites, frameworks, and bug hunts into a portfolio recruiters actually open, with case studies, linkable repos, live demo suites, and writeups structured like incident reports.
npx skillsauth add PramodDutta/qaskills QA Project Portfolio BuilderInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
Use this skill when the user:
Screeners open one link for roughly thirty seconds. They find: a fork of a tutorial repo, ten empty example tests against a todo app, no README, no CI, last commit eight months ago. That portfolio subtracts credibility.
A working portfolio is three artifacts, each complete, each answering a hiring question:
| Artifact | Hiring question it answers | |---|---| | A framework repo with CI | Can they build automation others could use? | | A bug-hunt writeup | Can they actually find problems? | | A test-strategy document | Can they think beyond individual tests? |
Three complete artifacts beat ten fragments. Depth reads as competence; breadth reads as tutorials.
Target a real public application (a demo e-commerce site, a public API), then build:
repo/
README.md <- the sales page (see below)
playwright.config.ts <- environments, projects, retries explained in comments
tests/
e2e/ <- 10-15 tests on the critical paths, POM structure
api/ <- API layer tests, shared auth/data setup
pages/ <- page objects with locator strategy visible
fixtures/ <- auth state, test data factories
.github/workflows/ <- CI running on push, badge green
README structure that converts:
npm ci && npx playwright test)The trade-offs section does the most work; it reads as engineering judgment, which is the thing being hired.
Pick a real public product, spend two focused hours, document 3-5 genuine findings:
Publish as a blog post, a repo markdown, or a PDF. Rules: no security probing beyond normal usage, no public shaming (redact or notify where appropriate), no fabricated findings. One real payment-flow papercut outweighs ten cosmetic nits.
For an imagined (or rebuilt) product, write 2-3 pages: scope and risks, what gets automated at which layer, what stays manual/exploratory, environments and data, entry/exit criteria, flake policy, tooling with justification. This is the document that separates senior applications; almost no candidate has one.
You cannot publish your employer's suite. You can rebuild the PATTERN:
Never copy code, config, or data from employer systems. The pattern is yours; the artifacts are theirs.
From zero, in order, each 1-2 weeks part-time:
Ship each fully before starting the next; an unfinished portfolio project is invisible, a finished small one is proof.
Recency is part of the proof: a green CI badge dated this week says maintained; the same badge dated last year says abandoned. Schedule the CI weekly (a cron trigger costs nothing) so the portfolio keeps proving itself while you job-hunt. And keep exactly one AI note honest: if agents helped generate tests, say so in the README and describe your review process; hiring teams now check for uncritically pasted AI output, and disclosed-plus-reviewed reads as modern practice.
development
Research market rates, build negotiation strategy, and create counter-offer scripts
data-ai
Track different resume versions, maintain master resume, manage tailored versions
testing
Customize resume for specific job postings while maintaining truthfulness
development
Create targeted resume sections optimized for different experience levels and roles