check-qa

Perform a structured QA pass on current changes or a feature, testing happy path, edge cases, and regression risks.

SKILL.md

GitHub

[Orchestrator Instructions] Do NOT execute this skill yourself. You MUST use the invoke_subagent tool to spawn an independent subagent with the Role: QA Automation Engineer.

QA target: $ARGUMENTS (feature name, PR number, or file path — leave blank to QA current changes)

This is the single-agent QA pass. For a multi-agent QA workflow (parallel functional / edge-case / regression analysts, consolidated report), use workflow-qa. That workflow drives this skill.

Gather what's being tested:

git diff main...HEAD --stat 2>/dev/null || git diff --stat HEAD~1
git log --oneline -5

Run a structured QA pass:

1. Acceptance criteria check

  • What is this feature/fix supposed to do? List the expected behaviors.
  • Verify each one is actually implemented (read the code, run if possible).
  • Flag any acceptance criteria that are missing or only partially implemented.

2. Happy path

Walk through the primary use case step by step. For each step:

  • What input/action?
  • What is the expected output/state?
  • Does the code produce that? (Check logic, not just syntax.)

3. Systematic Edge Cases & Boundaries

Utilize advanced matrix testing. Mark each ✅ pass / ❌ fail / ⚠️ untested:

Category Examples
Empty / null empty string, null/undefined, empty array/object
Boundary min/max values, off-by-one, $Min-1, $Max+1, max length
Invalid / Format wrong data type, negative numbers, invalid regex
State Transition test all valid and invalid state transitions
Concurrency race conditions, double-submit, stale data
HTTP/API Routing (BE only) explicit checks for accurate 400, 401, 403, 404, 500 status codes
Error states network failure, timeout, invalid input from external source

4. Automation Triaging & Manual Handoff

Evaluate if a test case is deterministic.

  • Level 1 (Fully Auto): Clear oracle, deterministic (e.g., math, formatting, data mapping). Proceed to test it.
  • Level 2 & 3 (Manual Handoff): Highly subjective (UI visual diffs), timing-dependent, or requires third-party black-box interaction. Do NOT attempt to run these automatically. Mark them as HANDOFF and append them to the Manual Testing List for the user.

5. Regression risk

List 3-5 existing features most likely to break from this change. Check each one in the code.

6. QA report

Write your report to .harness/reports/qa-report.md. Your chat response must ONLY be the path to this file (e.g. "QA complete. See .harness/reports/qa-report.md"). Do NOT output the report text in the chat. Format for the file:

## QA Report — <feature>

### ✅ Passing
- ...

### ❌ Failing
- <issue>: <expected> vs <actual>

### ⚠️ Needs manual testing
- ...

### Verdict
PASS / FAIL / CONDITIONAL PASS (needs fix before merge)