# QA Review

> Independent exploratory QA. Reads documentation, stands up the app, traverses UI workflows with agent-browser, finds edge cases. Use anytime to test app quality.

- Skill: `upsolve-labs/qa-review` (Agent Skill)
- Install (CLI): `npx skillmds@latest add upsolve-labs/qa-review`
- Raw SKILL.md: https://api.skillmd.com/api/skills/upsolve-labs/qa-review/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: AI & ML
- Author: Upsolve-Labs (https://skillmd.com/u/upsolve-labs)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/upsolve-labs/qa-review

---


## Update Check (run first)

```bash
_UPD=$(~/.claude/skills/upstack/bin/upstack-update-check 2>/dev/null || .claude/skills/upstack/bin/upstack-update-check 2>/dev/null || true)
[ -n "$_UPD" ] && echo "$_UPD" || true
```

If output shows `UPGRADE_AVAILABLE <old> <new>`: read `~/.claude/skills/upstack/skills/upgrade/SKILL.md` and follow the "Inline upgrade flow" (auto-upgrade if configured, otherwise AskUserQuestion with 4 options, write snooze state if declined). If `JUST_UPGRADED <from> <to>`: tell user "Running upstack v{to} (just updated!)" and continue.

# QA Review

You are performing independent exploratory QA. This is not tied to a specific branch or feature — you are testing the app as a user would.

## Step 1: Find Documentation
Read existing end-user documentation (README, docs/, wiki, etc).
If you cannot find it: AskUserQuestion for the documentation location.

## Step 2: Stand Up the App
Start a local instance of the app, or connect to one the user provides.
If you need auth or environment setup: AskUserQuestion.

## Step 3: Discover Flows
Identify all documented workflows and user-facing flows.
If you found too many flows or want to confirm scope: AskUserQuestion with the list of flows you found. Let the user select which to test.

## Step 4: Test Each Flow
For each selected flow, use `agent-browser`:

**Happy path:**
- `agent-browser open <url>` — navigate to starting point
- `agent-browser snapshot -i` — read the page structure
- Walk through the flow step by step, clicking @refs, filling forms
- `agent-browser screenshot evidence/qa/<flow-name>-step-N.png` at key states

**Edge cases:**
- Empty inputs, boundary values
- Rapid repeated actions, back button, refresh mid-flow
- Missing permissions, expired sessions
- Unexpected data shapes

**Note every issue found with:**
- What happened vs what was expected
- Screenshot of the broken state
- Steps to reproduce

## Step 5: Report
Summarize findings in an MD report with annotated screenshots.

## Step 6: Log Issues
List all findings to the user. AskUserQuestion: "Which of these should I log as TODOs?" Present each finding as a numbered option. Only create TODOS.md entries (and Linear tickets if MCP available) for the ones the user approves.