skillZs
★ LIVE SKILL TAGS ★
>>> LIVE SKILLS INDEX <<<
* OPEN SOURCE *
NO LOGIN, NO TRACKING
※ REAL INSTALL DATA ※
← back to all skills
comet-ml/opik50 installs

playwright-e2e

Playwright E2E test generation workflow for Opik. Use when generating, fixing, or planning automated tests in tests_end_to_end/.

How do I install this agent skill?

npx skills add https://github.com/comet-ml/opik --skill playwright-e2e
view source ↗

Is this agent skill safe to install?

  • Gen Agent Trust Hubwarn

    The skill provides an automated Playwright testing workflow for the Opik application but includes high-risk instructions to programmatically modify the user's home directory configuration and the application's React source code. The instructions to operate autonomously without user confirmation increase the risk of unauthorized code injection.

  • Socketpass

    No alerts

  • Snykpass

    Risk: LOW · No issues

  • Runlayerpass

    5/8 files flagged

What does this agent skill do?

Playwright E2E Test Workflow

Three-agent workflow for generating end-to-end tests for the Opik application:

  1. Planner -> Explores the Opik UI, generates a markdown test plan with fixture/page-object references
  2. Generator -> Transforms the plan into executable Playwright tests following Opik conventions
  3. Healer -> Automatically fixes failing tests with Opik-specific debugging knowledge

When to Use

Use when developer requests E2E test generation or repair:

  • "Generate an E2E test for the new dashboard"
  • "Create automated test for the upload flow"
  • "Add happy path test for experiment creation"
  • "Fix the failing prompts test"

Prerequisites

# Opik must be running locally
# Frontend at http://localhost:5173
# Backend healthy

# Playwright environment
cd tests_end_to_end/typescript-tests
npm install
npx playwright install chromium

Safety: Verify Local Config

Before running any tests or SDK operations, check ~/.opik.config. The Python Opik SDK (used by the Flask test helper service) reads this file. If it points to a cloud environment, the workflow will accidentally create real data there.

cat ~/.opik.config

If url_override points to anything other than http://localhost:5173/api, back up and fix it:

cp ~/.opik.config ~/.opik.config.bak 2>/dev/null || true
cat > ~/.opik.config << 'EOF'
[opik]
url_override = http://localhost:5173/api
workspace = default
EOF

After the workflow is complete, remind the user to restore their config: cp ~/.opik.config.bak ~/.opik.config

Knowledge Base

These context files provide the domain knowledge agents need to produce accurate tests:

Workflow Phases

Phase 1: Planning

Agent: agents/playwright-test-planner.md Input: Running Opik app + feature description Output: tests_end_to_end/typescript-tests/specs/{feature-name}.md

Uses seed test: tests/seed-for-planner.spec.ts

Phase 2: Generation

Agent: agents/playwright-test-generator.md Input: Markdown test plan from specs/ Output: tests_end_to_end/typescript-tests/tests/{feature-area}/{test-name}.spec.ts

Uses existing fixtures and page objects. References test-conventions.md for all coding standards.

Phase 3: Healing

Agent: agents/playwright-test-healer.md Input: Failing test + error info Output: Passing test or test.fixme() if the feature is genuinely broken

Directory Structure

tests_end_to_end/
├── test-helper-service/        # Flask service bridging TS tests to Python SDK
├── test_files/                 # Test attachments (images, audio, PDFs)
├── installer_utils/            # Shell scripts for environment checks
└── typescript-tests/
    ├── specs/                  # Markdown test plans (planner output)
    ├── tests/                  # Executable tests (generator output)
    │   ├── projects/
    │   ├── datasets/
    │   ├── experiments/
    │   ├── prompts/
    │   ├── tracing/
    │   ├── feedback-scores/
    │   ├── playground/
    │   ├── online-scoring/
    │   └── seed-for-planner.spec.ts
    ├── fixtures/               # Test fixtures (base -> feature-specific)
    ├── page-objects/           # Page Object Model classes
    ├── helpers/                # TestHelperClient, random, wait utilities
    ├── config/                 # Environment configuration
    └── playwright.config.ts    # Playwright configuration

Add the canonical catalog link to the repository README so users can inspect current installs and available audits. The publishing guide covers the complete discovery path.

<a href="https://skillzs.dev/skills/comet-ml/opik/playwright-e2e">View playwright-e2e on skillZs</a>