Stats
Actions
Tags
From skills
Writes and enhances the higher test layers — end-to-end (Playwright), API/contract, performance/benchmark, and accessibility tests. Use this whenever a feature needs browser, integration, api, perf, or a11y coverage, when the user asks to add or extend e2e/api/perf tests, or after exploratory QA surfaces a behavior worth locking in. For unit/seam tests written test-first, use practicing-tdd instead.
How this skill is triggered — by the user, by Claude, or both
Slash command
/skills:authoring-testsThis skill is limited to the following tools:
The summary Claude sees in its skill listing — used to decide when to auto-load this skill
These are the layers that prove the system works the way a user (or a caller) actually
These are the layers that prove the system works the way a user (or a caller) actually hits it — through a real browser, over HTTP, under load, with assistive tech. They catch what unit tests can't: htmx swaps, the live SSE transport, focus/keyboard behavior, wire contracts, and latency.
| Layer | Proves | Lives in |
|---|---|---|
| e2e | real browser flows, htmx swaps, SSE | e2e/tests/*.spec.ts |
| api/contract | content-types, escaping/XSS, cookies, isolation, malformed-payload tolerance | internal/web/api_test.go |
| perf | render/reducer hot paths; page latency | internal/web/bench_test.go + Playwright perf |
| a11y | WCAG 2.1/2.2 A/AA via axe-core | e2e/tests/a11y.spec.ts |
| See references/test-layers.md for what belongs where. |
Don't hand-write brittle selectors. Follow the Playwright agent workflow (the bundled
playwright-cli skill provides the tooling): explore → write a spec → generate the test →
heal until green. See references/playwright-agents.md.
scripts/init-agents.sh # one-time: npx playwright init-agents --loop=claude
scripts/run-layer.sh e2e # build demo + run a layer (e2e|api|a11y|ux|perf)
Prefer role and test-id locators and web-first assertions (expect(locator).toBeVisible())
over CSS/text and over waitForTimeout. Hard waits are the main reason e2e suites flake;
the framework's standards live in governing-qa-framework.
The browser suite runs against one shared in-memory demo session (workers: 1). Per-session
counters accumulate across the whole run, so assert relatively (read → act → assert it
increased), never a fixed value, and seed anything a test drives (forge rows, models) in
-demo mode — don't assume it's on disk. See references/demo-gotchas.md.
This skill writes the higher layers. practicing-tdd owns unit/seam; hardening-tests
strengthens any suite; governing-qa-framework owns the harness config/standards, not the
individual tests. When exploratory QA finds a bug, this is where the keeper regression lands.
npx claudepluginhub dotts-h/claude-skills --plugin skillsProvides CDSS development patterns for drug interaction checking, dose validation, clinical scoring (NEWS2, qSOFA), and alert classification integrated into EMR workflows.