跳到正文
cccskills
免费在 GitHub 公开

playwright-demo

Record narrated demo videos of the Rossoctl platform using Playwright and OpenAI TTS

中文简介正在准备中。当前显示原文说明。

rossoctl/rossoctl星标 303(GitHub 303 + cccskills 0)
集成工具PlaywrightOpenAI

实测评分 未实测

查看安装方法

包含的文件(1)

  • SKILL.md5.2 KB

SKILL.md(原文)

安装前可以查看交给 AI 代理的指令内容。

Playwright Demo Recording

Record demo videos of the Rossoctl platform with AI-generated voiceover narration.

When to Use

  • Creating a demo video of the Rossoctl platform for presentations
  • Recording a walkthrough of a new feature with voiceover
  • Generating reproducible video documentation of platform workflows

Prerequisites

  • HyperShift or Kind cluster with Rossoctl deployed
  • The playwright-demos worktree: .worktrees/playwright-demos/
  • ffmpeg and ffprobe installed
  • OPENAI_API_KEY in .env for narration (optional -- video-only without it)

Quick Start

cd .worktrees/playwright-demos

List available tests:

./local_experiments/run-playwright-demo.sh --cluster-suffix <SUFFIX>

Record video only:

./local_experiments/run-playwright-demo.sh --cluster-suffix <SUFFIX> --test <name>

Record video + narration:

source .env
./local_experiments/run-playwright-demo.sh --cluster-suffix <SUFFIX> --test <name>

Workflow

On invocation

  1. Write/review test using test:playwright -- verify markStep segments, assertions, selectors, and narration alignment
  2. Run test to verify it passes: --no-narration first
  3. Record video with voiceover: source .env then run
  4. Validate alignment and iterate on narration text
  5. Fix failures using playwright-demo:debug if needed

Recording pipeline (with OPENAI_API_KEY)

Step 1: Fast run -> measure video slot timing per markStep()
Step 2: Generate TTS per [section], inject waitForTimeout() to match
Step 3: Record narration-synced video -> composite voiceover

Test Review Requirements

Before recording, verify every test meets these criteria:

Section time slots

  • Every markStep() call is OUTSIDE conditional blocks
  • Each section is 3-15 seconds (split longer ones)
  • markStep() names match [section] in narration file 1:1

Assertive tests (fail if content missing)

Tests MUST use expect() assertions, NOT silent .catch(() => {}):

// REQUIRED - fails if navigation broken
await expect(page).toHaveURL(/\/agents/, { timeout: 10000 });

// REQUIRED - fails if content missing
await expect(element).toBeVisible({ timeout: 5000 });
expect(count).toBeGreaterThan(0);

// REQUIRED - after login
expect(page.url()).not.toContain('/realms/');

Keep .catch() ONLY for truly optional elements (VERIFY_PROFILE, hover effects).

Verification

Run the test without narration first to confirm it passes:

./local_experiments/run-playwright-demo.sh --cluster-suffix <SUFFIX> --test <name> --no-narration

If it fails, fix selectors by reading UI source code. Use playwright-demo:debug.

Directory Structure

demos/
├── 01-demos/full-platform-walkthrough/
├── 02-ui-pages/{home-overview,agent-detail,...}/
├── 03-workflows/{agent-chat,multi-namespace,...}/
├── 04-observability/{mlflow-traces,kiali-mesh,...}/
└── 05-advanced/{keycloak-admin,env-vars,...}/

Each demo directory contains all artifacts:

demos/02-ui-pages/home-overview/
├── home-overview.spec.ts              # Test spec (copy)
├── home-overview.txt                  # Narration text (copy)
├── home-overview-timestamps.json      # Step timings
├── audio_segments/                    # Per-section TTS (cached by MD5)
│   ├── intro.mp3 + intro.hash
│   ├── login.mp3 + login.hash
│   └── ...
├── home-overview_YYYY-MM-DD.webm      # Timestamped raw video
├── home-overview_YYYY-MM-DD_voiceover.mp4
├── home-overview_latest.webm          # Latest (overwritten)
└── home-overview_latest_voiceover.mp4

Audio segments are cached by MD5 hash -- only changed narration text triggers TTS regeneration.

Creating New Demos

  1. Add to demo-map.json
  2. Create e2e/<name>.spec.ts with markStep(), expect(), demoClick()
  3. Create narrations/<name>.txt with [section] markers
  4. Run and iterate

See playwright-research for analyzing UI code to write selectors.

Google Drive Sync

./local_experiments/sync-to-gdrive.sh

Shows setup instructions for scoped service account. Then:

./local_experiments/sync-to-gdrive.sh --sync --latest-only

Troubleshooting

ProblemFix
Validation IDLE gapsAdd narration text to fill the gap
Narration overlaps--sync pipeline adds extra waits automatically
Missing markStep timestampMove markStep() outside conditional blocks
Auth lost after navigationUse SPA sidebar clicks, not page.goto()
Overlapping audioFix markStep() calls, check timestamps file
Video in wrong directoryAdd test to demo-map.json

Task Tracking

  1. TaskList -- check existing demo recording tasks
  2. TaskCreate: playwright-demos | <cluster> | <test> | <phase>
  3. TaskUpdate as iterations progress

Related Skills

  • playwright-demo:debug -- Debug failing Playwright steps
  • playwright-research -- Analyze UI code, detect changes, plan videos
  • hypershift:cluster -- Create/manage test clusters
  • k8s:health -- Verify platform health before recording

评价

还没有评价。欢迎分享你的使用感受。

评论

无需登录即可发表。欢迎分享使用感受、提问,或说明在哪些环境下无法运行。

还没有评论。来写第一条吧。

提交后会立即公开。请勿填写邮箱、电话等个人信息。URL 最多 2 个。

0 / 1000 字

此表单受 reCAPTCHA 保护,适用 Google 的隐私权政策和服务条款。

同一仓库的技能

auth

免费

Authentication and authorization patterns for Rossoctl services

中文简介正在准备中。当前显示原文说明。

rossoctl/rossoctl3032026年10月25日 更新

Create confidential OAuth2 clients in Keycloak for server-to-service authentication

中文简介正在准备中。当前显示原文说明。

rossoctl/rossoctl3032026年10月25日 更新

Configure MLflow with mlflow-oidc-auth plugin for Keycloak OIDC authentication

中文简介正在准备中。当前显示原文说明。

rossoctl/rossoctl3032026年10月25日 更新

Configure OpenTelemetry Collector with OAuth2 authentication for trace export

中文简介正在准备中。当前显示原文说明。

rossoctl/rossoctl3032026年10月25日 更新

ci

免费

CI pipeline monitoring, status checks, and PR validation workflows

中文简介正在准备中。当前显示原文说明。

rossoctl/rossoctl3032026年10月25日 更新

Monitor running CI jobs, wait for completion, create tasks for results

中文简介正在准备中。当前显示原文说明。

rossoctl/rossoctl3032026年10月25日 更新

查看 rossoctl 的全部技能

报告此技能的问题