argent-qa-flows

Creates repeatable QA regression E2E tests as Argent flows from test cases and acceptance criteria.

2.4k|96|Updated Feb 23, 2026
One-click install
npx skills add https://github.com/software-mansion/argent --skill argent-qa-flows
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: argent-qa-flows
Source: https://github.com/software-mansion/argent/tree/main/packages/skills/skills/argent-qa-flows
Command: npx skills add https://github.com/software-mansion/argent --skill argent-qa-flows

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Turning tickets, test cases, and acceptance criteria into repeatable automated regression tests is slow and error-prone, and ad-hoc UI checks rarely survive reruns. This Skill converts those requirements into deterministic Argent flows with stable selectors, executable evidence, and a two-pass completion gate.

Core Features & Use Cases

  • Test contract authoring: Maps every requirement to a hard await:, assert:, or reviewed snapshot: check with a compact contract table before touching the app.
  • Deterministic setup and teardown: Enforces a launch-first structure, seeded baselines, and end-state restoration so repeated runs do not accumulate artifacts.
  • Two-pass proof gate: Requires two consecutive green runs with the same runner, fresh mobile services on pass 1, and resolution of all warnings before completion.
  • Use Case: Given a ticket saying "selecting Dark theme persists and Light is absent", generate a qa-settings-dark-theme flow that records the scenario, proves the state change with structural checks and a snapshot, and passes twice in CI.

Quick Start

Ask the agent to create a QA regression flow from your ticket or acceptance criteria, for example: "Create a QA flow verifying that enabling dark mode in Settings persists after restart on the iOS simulator."

Frequently Asked Questions about argent-qa-flows

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

FAQPage Schema
How do I turn a ticket or acceptance criteria into an automated regression test?

Define a compact test contract mapping each requirement to an action and executable evidence, then record the scenario as an Argent flow. The flow must pass twice consecutively with the same runner before it counts as complete.

How do I write E2E tests for mobile apps with stable selectors?

Prefer accessibility ids for targets and reserve coordinate fallbacks for genuinely unlabeled elements. Every screen change needs a destination identity check followed by an idle readiness check to avoid flaky timing.

Does this support Apple TV and Android TV testing?

No, Apple TV and Android TV are unsupported because touch directives fail at the gesture layer there. Use argent-tv-interact for those platforms; Vega (Fire TV) is supported via recorded tv-remote steps.

Why does my QA flow need two consecutive passes?

A single green run can hide timing luck or leftover state. Pass 1 starts with fresh Argent services and pass 2 runs immediately after, proving the flow is deterministic and independent of environment warmth.

What should I do when a flow step fails or errors?

Distinguish environment errors (unreadable tree, unresolvable run target) from product regressions like a failed launch. Fix environment issues and rerun, but report genuine app regressions instead of weakening the check.