skills.rest
laionazeredo
laionazeredoCOMMUNITY
Explore all skills by this author →
GitHub

che-manual-test-executor

Executes manual test plans via Playwright MCP and HTTP requests, producing per-AC pass/fail reports with screenshot evidence.

Updated May 12, 2026
Search Tags:#browser automation#acceptance criteria#manual testing#playwright#qa automation#test evidence#pass fail report
One-click install
npx skills add https://github.com/laionazeredo/che-ai --skill che-manual-test-executor-laionazeredo
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: che-manual-test-executor
Source: https://github.com/laionazeredo/che-ai/tree/main/skills/che-manual-test-executor
Command: npx skills add https://github.com/laionazeredo/che-ai --skill che-manual-test-executor-laionazeredo

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Manual acceptance testing of a written test plan is slow and inconsistent: someone must open the browser, click through each acceptance criterion, capture screenshots, and record results. This Skill automates that execution loop, turning a manual_test_plan.md file into a structured, evidence-backed pass/fail report. ## Core Features & Use Cases - Step-by-step AC execution: Parses each AC block (GIVEN/WHEN/THEN plus numbered Manual Steps) and dispatches every step to Playwright MCP (navigate, click, fill, select, upload) or HTTP drivers (GET/POST/PUT/PATCH/DELETE). - Evidence capture per step: Saves PNG screenshots, visible-text snapshots, API response bodies, and browser console logs into a session-scoped evidence directory, plus a SHA256 evidence manifest with workspace-level thumbnails. - Structured reporting: Produces a per-AC verdict table (PASS/PARTIAL/FAIL/SKIP), smoke test results (build, lint, login, navigation, logs), human-review items, and an overall ship-readiness verdict. - Use Case: After che-act writes manual_test_plan.md for a checkout feature, run this Skill against the worktree to click through all five ACs in a real browser, capture screenshots of each step, and receive a report showing AC-3 failed because the success URL never loaded. ## Quick Start Execute the manual test plan for worktree /path/to/worktree with task-id T123 and generate the evidence-backed report.

Frequently Asked Questions about che-manual-test-executor

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I execute a manual test plan automatically with Playwright?▼

Provide the worktree root path and a task ID or explicit path to manual_test_plan.md. The Skill parses each AC block, dispatches every numbered step to Playwright MCP or HTTP drivers, captures screenshots per step, and returns a per-AC pass/fail report.

What is the difference between manual test execution and automated QA runs?▼

Automated QA runs shell commands like lint, typecheck, build, and unit or E2E tests with exit-code results. Manual test execution drives a real browser through the written manual steps, capturing screenshots, visible text, and API responses as evidence per acceptance criterion.

Can it test API-only scenarios without a browser?▼

Yes. Steps that mention curl or API calls are dispatched to HTTP drivers (GET, POST, PUT, PATCH, DELETE) instead of browser actions, and full response bodies are saved as JSON evidence files for each step.

What happens when a test step requires human verification?▼

Steps involving external systems like email sandboxes or server logs are marked HUMAN_REVIEW_NEEDED and skipped rather than faked. They are listed in the report's human-only items section for a reviewer to verify before merge.

Does it protect against testing production environments?▼

Yes. URLs matching production patterns are blocked and require double confirmation before proceeding. Credentials are never logged or screenshotted, and all evidence is written outside the user's worktree under session-scoped directories.

Related Skills

Explore More Software Engineering Skills→
LeviLeckenbyLeviLeckenby

spec

Generate Given/When/Then specifications with acceptance criteria for features.

Community
Advanced
JoyJoin-Tech-LimitedJoyJoin-Tech-Limited

harness-verification-gate

Validates plans and PR diffs against a 5-pillar engineering checklist and outputs a gate verdict.

Official
Intermediate
LimaTechnologiesLimaTechnologies

quality-gate

Run typecheck, lint, tests, and build for TypeScript projects.

Community
Advanced
LambdaTestLambdaTest
350

test-framework-migration-skill

Migrate test automation scripts between Selenium, Playwright, Puppeteer, and Cypress.

Official
Advanced