qa-level-assessment_skill

Overview

This skill evaluates QA test level for a change by combining story descriptions, code diffs, and conversational inputs to produce a risk-based QA recommendation. It outputs a numeric score, a risk level, and concrete guidance on who should test, who should review, and whether double-checks are required.

How this skill works

The skill collects required inputs (change summary, PR or diffs, changed files) and asks follow-up questions for missing details. It scores defect likelihood (1–3) and adds weighted risk factors (payment, PII, external integrations, DB schema, auth, broad impact). The final score maps to an explicit QA level and a recommended test workflow.

When to use it

• When you need to decide QA level for a PR before merging
• When determining whether a separate tester or double-check is required
• When setting the depth of test-case review or assigning QA resources
• When triaging changes that touch payments, PII, DB schema, or external systems
• When you want a repeatable, auditable QA recommendation for release planning

Best practices

• Provide a concise change summary and link to PR or paste diffs for accurate scoring
• Answer the risk checklist explicitly (payments, PII, external systems, DB, auth, scope)
• Use the numeric score to align testing effort with release windows and rollback plans
• Treat recommendations as minimums—escalate to QA lead for ambiguous or high-impact cases
• Document acceptance criteria and critical test cases alongside the QA recommendation

Example use cases

• A small UI copy change: expect low score and minimal testing by a non-author tester
• An API logic change touching multiple modules: expect medium score and a separate reviewer
• A payment flow change: expect high score, QA-assigned tests, and mandatory double-check
• A DB schema migration: expect elevated score and a plan for data integrity and rollback
• A change integrating a third-party service: include external-failure scenarios in testing