UX QA Gate

v1.0.0

Self-review gate for UI/UX work before delivering to the user. Run automatically after building, modifying, or completing any user-facing feature, page, comp...

1· 181·0 current·0 all-time

Install

OpenClaw Prompt Flow

Install with OpenClaw

Best for remote or guided setup. Copy the exact prompt, then paste it into OpenClaw for michellebolander/ux-qa-gate.

Previewing Install & Setup.
Prompt PreviewInstall & Setup
Install the skill "UX QA Gate" (michellebolander/ux-qa-gate) from ClawHub.
Skill page: https://clawhub.ai/michellebolander/ux-qa-gate
Keep the work scoped to this skill only.
After install, inspect the skill metadata and help me finish setup.
Use only the metadata you can verify from ClawHub; do not invent missing requirements.
Ask before making any broader environment changes.

Command Line

CLI Commands

Use the direct CLI path if you want to install manually and keep every step visible.

OpenClaw CLI

Bare skill slug

openclaw skills install ux-qa-gate

ClawHub CLI

Package manager switcher

npx clawhub@latest install ux-qa-gate
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Benign
high confidence
Purpose & Capability
Name/description match the runtime instructions. The skill is instruction-only, requests no binaries, env vars, or installs, and everything it asks for (checking UI/UX heuristics, functional checks, state sweeps) is appropriate for a QA gate.
Instruction Scope
SKILL.md confines actions to reviewing the work, re-reading the original request, and using the bundled heuristics checklist file. It does not instruct the agent to read unrelated system files, call external endpoints, or exfiltrate data. Note: the gate assumes the agent can interact with the UI or project artifacts to perform checks (manual or automated), but it does not add any instructions that would expand privilege or access beyond that scope.
Install Mechanism
No install spec and no code files — lowest-risk model. Nothing will be written to disk or downloaded by the skill itself.
Credentials
The skill requests no environment variables, credentials, or config paths. All required artifacts are bundled (heuristics checklist) or are the user's project/request context, which is appropriate for a QA gate.
Persistence & Privilege
always is false and model invocation is normal. The skill does not request persistent system presence or modify other skills or system-wide settings.
Assessment
This skill is coherent and low-risk: it is a textual QA checklist and asks for no credentials or installs. Before using it, confirm the agent already has the necessary access to your project or running app (build artifacts, staging site, or screenshots) so checks are meaningful. If you plan to run this gate in automation/CI, ensure the CI environment gives only the minimal required access (build artifacts, staging URL) and review any QA outputs before sharing externally, as the gate will re-read the original request and inspect the work you point it at. If you expect automated interaction with the UI, note the skill contains no automation tooling — you will need to provide or permit those tools separately.

Like a lobster shell, security has layers — review code before you run it.

latestvk972txctjhw843jjff8w7b9mbn835rbw
181downloads
1stars
1versions
Updated 1mo ago
v1.0.0
MIT-0

UX QA Gate

Run this gate after completing any user-facing work. Do not deliver work to the user until this gate passes.

When to Run

  • After building or modifying any page, component, or flow
  • After completing a feature request
  • Before saying "done" or presenting work to the user

Process

1. Functional Completeness Check

Before examining UX, verify the feature actually works end-to-end:

  • Every button/link has a working handler
  • Forms submit and process data correctly
  • Navigation flows connect (no dead ends)
  • Data displays where expected (not placeholder or undefined)
  • The feature does what was asked — re-read the original request

2. Heuristic Review

Walk through each of Nielsen's 10 heuristics against the work just completed. Load references/heuristics-checklist.md and evaluate each applicable item. Not every heuristic applies to every change — skip items that are clearly irrelevant to the scope of work, but err on the side of checking.

3. State & Edge Case Sweep

Check these states exist and are handled for every data-driven view:

StateWhat to verify
EmptyNo data yet — helpful message, not blank screen
LoadingSpinner, skeleton, or indicator while fetching
ErrorClear message with recovery path (retry, go back)
SuccessConfirmation after actions (toast, redirect, message)
PartialSome data missing — graceful degradation, not crash
OverflowLong text, many items, large images handled
AuthProtected routes redirect or show appropriate state

4. Interaction & Responsiveness

  • Click every interactive element — does it respond?
  • Tab through focusable elements — logical order?
  • Resize viewport — does layout hold at mobile/tablet/desktop?
  • Check touch targets on mobile — minimum 44×44px

5. Severity Classification

For each issue found:

SeverityMeaningAction
🔴 BlockerUser cannot complete the taskFix before delivering
🟠 MajorUser struggles significantlyFix before delivering
🟡 MinorUser notices but can work aroundNote to user, fix if quick
🔵 EnhancementPolish or improvementNote to user for backlog

6. Decision

  • Any 🔴 or 🟠 issues → Fix them, then re-run the gate
  • Only 🟡 or 🔵 → Deliver the work with a brief note of known issues
  • No issues → Deliver confidently

Output Format

When reporting to the user, keep it brief:

✅ QA Gate passed — [summary of what was checked]

Or if issues were found and fixed:

🔧 QA Gate — found and fixed [N] issues:
- [brief description of each fix]

Or if minor issues remain:

⚠️ QA Gate — [N] minor items noted:
- [brief description + severity]

Do not produce a full heuristic evaluation report unless the user asks for one. The gate is a self-check, not a deliverable.

Comments

Loading comments...