تشغيل أي مهارة في Manus بنقرة واحدة

headless-browser-test

النجوم٠

التفرعات٠

آخر تحديث٢٦ يوليو ٢٠٢٦ في ١٩:١٠

Run headless browser smoke and E2E tests for a web service. Use when verifying rendered web UI, checking browser errors, creating Playwright-style scenarios, collecting screenshots, or testing a deployed or local service through Chromium.

التثبيت

التثبيت باستخدام Codex أو Claude انسخ هذا Prompt والصقه في Codex أو Claude أو مساعد آخر ليراجع صفحة Skill ويثبّتها لك.

تشغيل في Manus

المصدر

hikaMaeng

hikaMaeng/skills

فتح مستودع GitHub عرض مستودعات المنشئ

تنزيل

تشغيل في Manus

المهن ذات الصلةSOC

استنادا إلى تصنيف SOC المهني

محللو ضمان جودة البرمجيات والمختبرونمهن الحاسوب والرياضيات·SOC 15-1253

مستكشف الملفات

7 ملفات

SKILL.md

readonly

name	headless-browser-test
description	Run headless browser smoke and E2E tests for a web service. Use when verifying rendered web UI, checking browser errors, creating Playwright-style scenarios, collecting screenshots, or testing a deployed or local service through Chromium.

Headless Browser Test

Use the scripts first. Keep prompts short; let them enforce browser setup, artifacts, trace capture, and report shape.

The target URL is an input to this skill, never something it guesses or rewrites. Resolve a URL that is reachable from wherever the browser process runs, prove it is reachable, then test it.

Default Flow

Resolve the target URL from the user, the deploy report, service docs, or the local dev server output.
Prove the URL is reachable from the same environment that will run the browser:

curl -fsS <url> >/dev/null

Run a smoke test:

node .codex/skills/headless-browser-test/scripts/run-headless-browser-test.mjs --url <url>

For real E2E, write a small JSON scenario and run it with --spec.
If the user asks for durable repository tests, generate a Playwright template and fill only the marked TODOs:

sh .codex/skills/headless-browser-test/scripts/write-playwright-e2e-template.sh test/e2e

Load Only What You Need

references/environment.md: runtime requirements, missing Chromium or Playwright handling, and choosing a reachable URL.
references/scenario-authoring.md: scenario JSON format, stable locator contract, and fill-in templates for common tests.
references/operation.md: testing a deployed service, artifact paths, failure triage, and final-answer requirements.

Hard Rules

Resolve the browser and Playwright from the environment. If the runtime provides them globally, point HEADLESS_BROWSER_EXECUTABLE and HEADLESS_BROWSER_PLAYWRIGHT_ROOT at them rather than adding project dependencies for one-off browser evidence.
A missing browser in a container image is an image problem. Fix the image; do not add a workspace dependency to work around it.
When the browser runs somewhere other than the host that published the port, pass the address that resolves from the browser's side. Note that a page served from localhost is a secure origin and one served from another hostname is not, so APIs such as Web Crypto can behave differently — choose the URL deliberately and record which one was tested.
Test rendered behavior a user can see or operate. Do not assert implementation details.
Prefer role/name, label, text, alt text, title, and test id locators. Use CSS or XPath only when no stable user-facing or documented test hook exists.
Keep tests isolated. Do not rely on execution order, shared browser storage, or state left by a previous test.
Avoid uncontrolled third-party pages and services. Mock or constrain them when they affect the result.
Prefer auto-waiting locator actions and web-first assertions over manual sleeps or one-shot visibility checks.
Capture evidence under test/YYYYMMDD/<HHMMSS>_headless-browser-test/.
Final answers must include pass/fail, the tested URL, the report path, screenshots, and whether the run was smoke-only or scenario-based.

المزيد من هذا المستودع

نفس المستودع

docker-compose-module-design

hikaMaeng/skills

Enforce the repository Docker contract: Docker-first runtime, one root docker-compose.yml, module-owned Docker assets, apps service-module to container mapping, isolated networking, Docker socket passthrough for container-spawning services, a standard npm deploy entrypoint, and runtime-only Dockerfiles that copy prebuilt local dist artifacts instead of building the project. Use when defining compose services, Dockerfiles, networks, volumes, env_file wiring, deploy scripts, Docker socket access, or container ownership.

2026-07-260

i18n-resource-map

hikaMaeng/skills

Enforce that every user-visible string comes from a translated resource key, never a literal in the view. Covers the flat dotted-key resource files per language, the required language set, and the per-folder static key map that keeps a component's resource surface small. Use when writing or reviewing any view, component, label, message, aria-label, or user-facing text.

2026-07-260

monorepo-architecture-guard

hikaMaeng/skills

Enforce the repository monorepo contract: TypeScript latest stable, Turbo plus npm workspaces, exact versions, repository package reuse, package and app boundaries, service-module shape, dist manifest targets, env-schema usage, deploy-script conventions, and React plus shadcn/ui plus Express wiring. Use when creating or reviewing packages, app service modules, package.json, tsconfig, workspace imports, env handling, frontend and backend integration, or monorepo configuration.

2026-07-260

package-docs-graph

hikaMaeng/skills

Enforce the repository package documentation contract as a navigable dependency graph. Use when creating or updating package README, docs pages, navigation tables, architecture router, API docs, usage docs, constraints, blast-radius/consumer records, internals, decisions, or testing docs. Docs exist so a later model drills down from task to edit target with minimal context and sees blast radius before editing. Keep content terse; keep only repo-specific contracts, paths, edges, intent, and decisions.

2026-07-260

react-model-render

hikaMaeng/skills

Use when designing, implementing, reviewing, or refactoring a React front end. Enforces Model Render — the screen is a projection of a pure in-memory model that lives outside React; React state is only an update trigger (never the source of truth); each component subscribes to the smallest model slice it reads. Prevents state-as-source-of-truth coupling, immutable-copy churn, lifecycle leaks, root/global-state ownership, prop/listener meshes, poor partial rendering, and fake refactors that only relocate complexity.

2026-07-260

web-service-scaffold

hikaMaeng/skills

Use only when explicitly invoked as $web-service-scaffold to run the bundled web-service scaffold script.

2026-07-260

Headless Browser Test

Use the scripts first. Keep prompts short; let them enforce browser setup, artifacts, trace capture, and report shape.

The target URL is an input to this skill, never something it guesses or rewrites. Resolve a URL that is reachable from wherever the browser process runs, prove it is reachable, then test it.

Default Flow

Resolve the target URL from the user, the deploy report, service docs, or the local dev server output.

Prove the URL is reachable from the same environment that will run the browser:

curl -fsS <url> >/dev/null

Run a smoke test:

node .codex/skills/headless-browser-test/scripts/run-headless-browser-test.mjs --url <url>

For real E2E, write a small JSON scenario and run it with --spec.

If the user asks for durable repository tests, generate a Playwright template and fill only the marked TODOs:

sh .codex/skills/headless-browser-test/scripts/write-playwright-e2e-template.sh test/e2e

Load Only What You Need

references/environment.md: runtime requirements, missing Chromium or Playwright handling, and choosing a reachable URL.

references/scenario-authoring.md: scenario JSON format, stable locator contract, and fill-in templates for common tests.

references/operation.md: testing a deployed service, artifact paths, failure triage, and final-answer requirements.

Hard Rules

Resolve the browser and Playwright from the environment. If the runtime provides them globally, point HEADLESS_BROWSER_EXECUTABLE and HEADLESS_BROWSER_PLAYWRIGHT_ROOT at them rather than adding project dependencies for one-off browser evidence.

A missing browser in a container image is an image problem. Fix the image; do not add a workspace dependency to work around it.

When the browser runs somewhere other than the host that published the port, pass the address that resolves from the browser's side. Note that a page served from localhost is a secure origin and one served from another hostname is not, so APIs such as Web Crypto can behave differently — choose the URL deliberately and record which one was tested.

Test rendered behavior a user can see or operate. Do not assert implementation details.

Prefer role/name, label, text, alt text, title, and test id locators. Use CSS or XPath only when no stable user-facing or documented test hook exists.

Keep tests isolated. Do not rely on execution order, shared browser storage, or state left by a previous test.

Avoid uncontrolled third-party pages and services. Mock or constrain them when they affect the result.

Prefer auto-waiting locator actions and web-first assertions over manual sleeps or one-shot visibility checks.

Capture evidence under test/YYYYMMDD/<HHMMSS>_headless-browser-test/.

Final answers must include pass/fail, the tested URL, the report path, screenshots, and whether the run was smoke-only or scenario-based.