skills.rest
Yassimba
YassimbaCOMMUNITY
Explore all skills by this author →
GitHub

e2e-test

Drives live applications first-hand across CLI, API, and web UI surfaces to capture verifiable end-to-end evidence.

9|Updated Jul 17, 2026
Search Tags:#browser automation#manual testing#e2e testing#api testing#smoke test#bug hunt#datastore validation
One-click install
npx skills add https://github.com/Yassimba/loom --skill e2e-test-yassimba
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: e2e-test
Source: https://github.com/Yassimba/loom/tree/main/skills/e2e-test
Command: npx skills add https://github.com/Yassimba/loom --skill e2e-test-yassimba

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Automated test suites can pass while the real application is broken — synthetic harnesses miss startup ordering, real network paths, real datastore constraints, and real rendering. This Skill replaces hearsay with first-hand evidence by driving the live app directly and validating every claim against artifacts on disk and direct datastore queries. ## Core Features & Use Cases - Multi-surface walkthroughs: Discovers and exercises CLIs, APIs, web UIs, and datastores, capturing stdout, HTTP responses, screenshots, and query results as evidence. - Structured six-phase workflow: Pre-flight inventory, parallel research sub-agents, app startup, journey-by-journey testing with edge-case sweeps and bug probes, cleanup, and a final report. - Real bug handling: When a genuine defect is found, it documents expected vs actual behavior, writes a regression unit test, and re-verifies the fix first-hand. - Use Case: Before shipping a release, ask the agent to run a manual E2E walkthrough — it starts the server, clicks through the web UI, hits the API with curl, confirms database rows landed, and produces an evidence-backed report of issues found and fixed. ## Quick Start Use the e2e-test skill to manually walk through this app's CLI, API, and web UI and report any bugs with evidence.

Frequently Asked Questions about e2e-test

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

FAQPage Schema
How do I manually end-to-end test a web application with an AI agent?▼

Run a structured walkthrough that starts the live app, drives each surface (CLI via Bash, API via curl, web UI via agent-browser), and captures evidence files for every step. Each claim is verified by reading back artifacts and querying the datastore directly.

Why is a passing test suite not enough before release?▼

Test suites use synthetic harnesses like TestClient, fixtures, and mocked browsers that miss real-world failures: startup ordering, real network paths, actual datastore constraints, real auth, and real rendering. Driving the live app first-hand exposes what breaks in production.

What tools are needed for browser-based E2E walkthroughs?▼

Web UI testing uses agent-browser, which requires Linux, WSL, or macOS and is installed via npm. It supports clicking, form filling, screenshots, viewport resizing for responsive checks, and console error inspection.

How do I verify database side effects during manual testing?▼

Query the datastore directly after every state-changing action using the project's actual client — psql, mysql, sqlite3, mongosh, redis-cli, aws s3 ls, or kafkacat. Confirm created rows, updated fields, deletions, cascades, and the absence of orphaned records.

What happens when a real bug is found during the walkthrough?▼

The bug is documented with expected vs actual behavior and evidence file paths, then a focused regression unit test is written in the existing suite. After the fix goes green, the failing E2E step is re-driven first-hand with fresh evidence.

When should I not use manual E2E walkthroughs?▼

Skip it when the project has no user-facing interfaces — no CLI, API, web UI, or observable datastore side effects — since there is nothing to drive. It also does not replace unit or integration tests; it complements them with first-hand verification.

Related Skills

Explore More Software Engineering Skills→
AlabamaMikeAlabamaMike

using-git-worktrees

Create isolated Git worktrees with safety verification and dependency setup.

Community
Intermediate
stevenmunozstevenmunoz

fix-pr-review

Triage and fix Claude Code Review PR comments with minimal changes.

Community
Advanced
JunnnnyWonJunnnnyWon

smoke-check

Run the automated test suite and generate a PASS/FAIL smoke-test report.

Community
Intermediate
simotasimota
68

Voyager

Builds end-to-end browser automation suites with Playwright, Cypress, or WebdriverIO.

Community
Advanced