---

name: qa description: QA workflow for software projects. Use for tasks like writing test plans, deriving test cases from PRDs, risk-based testing, API testing, UI testing, test data design, defect reporting, release checklists, and setting up automation strategy (unit/integration/e2e) with clear entry/exit criteria.

qa

Use this skill to drive Quality Assurance work end-to-end: test planning, test case design, execution strategy, and release readiness.

Core outputs (pick what the task needs)

• Test plan (scope, approach, environments, entry/exit criteria)
• Risk matrix (severity × probability × detectability)
• Test cases (functional, negative, boundary, compatibility)
• API test checklist (contracts, auth, idempotency, error codes)
• Regression suite definition and release checklist
• Automation strategy (what to automate, where to stop)

Workflow

1. Understand the product and changes

• Identify user journeys, critical flows, and non-functional requirements.
• List assumptions and external dependencies (3rd party APIs, DB migrations).

1. Risk-based test design

• Classify features by business criticality and change frequency.
• Prioritize tests for high-risk areas: auth, payments, data integrity, permissions, migrations.

1. Derive test cases

• Happy path + negative path + boundary conditions.
• State transitions, concurrency, retries, and idempotency for APIs.
• Localization/timezone, accessibility, performance where relevant.

1. Test data and environment

• Define test accounts/roles.
• Seed data strategy; ensure repeatability and cleanup.
• Identify environment parity gaps (staging vs prod).

1. Execution and reporting

• Run smoke → functional → regression.
• Record evidence, steps to reproduce, expected vs actual, logs/screenshots.
• Triage and retest; track flaky issues explicitly.

1. Release readiness

• Define entry/exit criteria.
• Prepare rollback plan and monitoring checks (error rate, latency, core KPIs).

Defect report template (use verbatim)

• Title:
• Environment:
• Version/commit:
• Severity / Priority:
• Preconditions:
• Steps to reproduce:
• Expected:
• Actual:
• Attachments (logs/screenshots/video):
• Notes (suspected area, frequency, impact):

Automation guidance (practical)

• Automate high-value stable paths first (smoke + core regression).
• Prefer unit/integration tests before E2E.
• Keep E2E small, deterministic, and focused on journeys.
• Add contract tests for APIs; pin schemas and error codes.