product-review
Skill2 starsChanged 35 days agoMITView on GitHub
Produce a visual HTML review of a PRD — problem statement, success criteria, and HTML wireframe mockup — to align on what the user will see before designing architecture or writing code. Use after /to-spec and before /system-architecture.
11 other projects use a version of this.
Adopted by
Most recently taken up by
- victor-falcon/skillstheir own version, first seen Nov 8, 2026
- Bobagi/claude-skillstheir own version, first seen Oct 29, 2026
- jarrheyd/skillstheir own version, first seen Oct 11, 2026
- sendaifun/solana-newtheir own version, first seen Oct 9, 2026
- vishal-mandhane/divinitytheir own version, first seen Sep 25, 2026
- RizwanZafaris/product-manager-OStheir own version, first seen Sep 3, 2026
Other versions11 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 11
- sendaifun/solana-new★ 178same name
- victor-falcon/skills★ 16same name
- RizwanZafaris/product-manager-OS★ 4same name
- vishal-mandhane/divinity★ 3same name
- d-wwei/product-review★ 2same name
- Iteam-company/claude-code-boilerplates★ 2same name
- AlbertFD/AlbertFD-skills★ 1same name
- chawanjoey/claude-skills★ 1same name
- AlexZWANG1/Augur★ 0same name
- Bobagi/claude-skills★ 0same name
1 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository3 other files it gives its agents.
- implement-loopSkill
- program-designSkill
- system-architectureSkill
The file
We split this skill into the rules it gives an agent
11 rules. 0 are common, 6 are distinctive, and 5 are very specific to this project.
0
Common
Most of the 7 edited versions we read keep these.
6
Distinctive
This author's own choices. Worth a look.
5
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 31
Read existing UI patterns: If the project is a web app, explore existing pages, components, and layouts.
- Very specific: line 36
Read existing mockups: Check if mockups or wireframes already exist in a docs/ directory, a Figma link in the PRD, or a mockups/ folder.
- Distinctive: line 45
Not every PRD needs a visual mockup.
- Distinctive: line 111
Use only HTML + inline CSS.
- Distinctive: line 116
No AI-look defaults: avoid the generic AI-generated pattern set — default purple/indigo gradient heroes, centered glassmorphism cards, …
- Distinctive: line 123
If the change affects multiple screens/states (e.g., empty state, error state, confirmation), show each in a separate mockup panel with a …
- Very specific: line 125
Use border: 2px dashed #999 for wireframe-style containers.
- Distinctive: line 159
Multi-screen flows: if the feature spans multiple screens (wizard, onboarding), show them left-to-right as a sequence with arrow separators.
- Very specific: line 164
Write the complete HTML to .scratch/product-review/<feature-name>/index.html.
- Distinctive: line 166
Create the directory if it doesn't exist.
- Very specific: line 180
If the user requests changes, update the HTML and overwrite the file.
“Common” is measured: kept by at least 6 in 10 of the 7 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Product Review
Bridge the gap between product requirements (text in a PRD) and architecture decisions (how services communicate). Produce a visual spec — a self-contained HTML page showing the problem, the success criteria, and wireframe mockups of the UI — so everyone agrees on what the user sees and interacts with before any code is written.
$ARGUMENTS is the PRD issue to review. Pass a number, URL, or path.
The issue tracker should have been provided to you — run
/setup-matt-pocock-skills if not.
Process
1. Gather context
- Read the PRD. Fetch it from the tracker. Read the full body: problem statement, user stories, solution, implementation decisions, testing decisions, out of scope.
- Read existing UI patterns. If the project is a web app, explore existing
pages, components, and layouts. Look at
CONTEXT.mdfor domain vocabulary and UI conventions. Check for existing design system components, CSS variables, or layout patterns — your mockup should match the project's visual language. - Read existing mockups. Check if mockups or wireframes already exist in a
docs/directory, a Figma link in the PRD, or amockups/folder. If they exist, reference them and show only the diff.
Done when: you have read the PRD body, surveyed existing UI patterns, and identified whether mockups already exist.
2. Classify the PRD
Not every PRD needs a visual mockup. Before producing artifacts, classify:
| If the PRD… | Produce |
| Adds new UI (page, dialog, form, nav) | Full HTML: problem card + criteria + mockup |
| Modifies existing UI (new field, changed layout) | Mockup showing the delta (before/after or overlay) |
| Is API-only, data-only, or backend only | Skip mockup — produce only problem + criteria card |
| Changes user-facing behavior without new UI (copy, permissions, flow) | Problem card + criteria + user flow description as prose |
3. Produce the HTML page
Produce a single self-contained HTML file at
.scratch/product-review/<feature-name>/index.html. Inline all styles — no
external CSS, no JS, no framework. The file renders in any browser.
The page has three sections. Only the mockup section can be omitted (see §2).
Section A: Problem statement card
A prominent card at the top. Render the PRD's problem statement as a concise, readable block. Use the project's brand colors if you can infer them (check CSS variables in the codebase, or use a neutral palette).
<div class="card problem">
<h2>Problem</h2>
<p>As a field agent, I currently cannot update incident reports after
submission. New evidence collected on-site must be appended via email to
dispatch, causing delays and data-entry errors.</p>
</div>Section B: Success criteria checklist
Render each acceptance criterion as a checkbox list. Use the PRD's testing decisions and user stories to derive the list.
<div class="card criteria">
<h2>Success Criteria</h2>
<ul class="checklist">
<li><input type="checkbox" disabled checked> Agent can edit report within 24h of submission</li>
<li><input type="checkbox" disabled> Edits are logged as revision history</li>
<li><input type="checkbox" disabled checked> Manager receives notification on edit</li>
</ul>
</div>Rules:
- Derive criteria from user stories, testing decisions, and acceptance criteria in the PRD. If the PRD doesn't have explicit criteria, synthesise them from the stories and solution.
- Cover every distinct persona in the user stories — at least one criterion per persona, phrased so it can fail ("an admin can X" is checkable; "users love X" is not).
- Cover the edge-case states — empty, error, loading, long content, small viewport. A criterion that cannot fail is decoration.
- Mark any that are clearly already met by existing functionality as
checked. - Group related criteria visually with indentation or subtitles.
Section C: Wireframe mockup
An HTML+CSS representation of the UI. This is not a screenshot or an image — it's semantic HTML styled to look like a wireframe.
Rules:
- Use only HTML + inline CSS. No images, no SVGs, no JavaScript.
- Grey-scale wireframe style (different shades of grey for containers, borders,
text blocks, using
#e0e0e0,#f5f5f5,#333, etc.) — unless the project has clear brand colors, in which case use them for emphasis. - Lorum ipsum or
[placeholder text]for content the PRD doesn't specify. - No AI-look defaults: avoid the generic AI-generated pattern set — default purple/indigo gradient heroes, centered glassmorphism cards, near-identical card rows, filler copy. Derive the palette and layout from the project's brand (CSS variables) or, when the project has none, from a reference site's design system; state which you used.
- Show realistic structure: real header, real nav items, real button labels from the PRD.
- If the change affects multiple screens/states (e.g., empty state, error state, confirmation), show each in a separate mockup panel with a label.
- Use
border: 2px dashed #999for wireframe-style containers. - The mockup should fill ~60-80% of the page width; center it.
<div class="mockup">
<h3>Screen: Edit Incident Report</h3>
<div class="wireframe">
<div class="wf-header">
<div class="wf-logo">[Logo]</div>
<div class="wf-nav">[Dashboard] [Reports] [Logout]</div>
</div>
<div class="wf-body">
<div class="wf-form">
<div class="wf-field"><label>Report ID</label><div class="wf-input">INC-2024-0421</div></div>
<div class="wf-field"><label>Description *</label><div class="wf-textarea">[Editable text area with existing content]</div></div>
<div class="wf-field"><label>Attachments</label><div class="wf-input">[Upload new evidence...]</div></div>
<div class="wf-actions">
<button class="wf-btn-primary">Save Edits</button>
<button class="wf-btn-secondary">Cancel</button>
</div>
</div>
</div>
</div>
</div>When to adapt fidelity:
| Situation | Style |
| This is a new greenfield feature | Detailed layout with fields, buttons, states |
| The PRD says "follow pattern of X" | Show the existing pattern, highlight the new piece |
| Only one new field on an existing page | Show the existing page with the new field highlighted (colored border or annotation) |
| Mobile / responsive concern | Show mobile viewport mockup (narrow container, 375px simulated width) |
Multi-screen flows: if the feature spans multiple screens (wizard, onboarding), show them left-to-right as a sequence with arrow separators.
4. Write the file
Write the complete HTML to .scratch/product-review/<feature-name>/index.html.
Create the directory if it doesn't exist.
The filename should match the feature name from the PRD title or ticket label — kebab-case, lowercase.
Tell the user where the file is:
"Product review ready at
.scratch/product-review/<feature-name>/index.html— open it in a browser to see the mockup.""Would you like to modify anything or proceed with system-architecture?"
5. Iterate on feedback
If the user requests changes, update the HTML and overwrite the file. Only
when they approve does the PRD move to system-architecture.
How this feeds other skills
| Skill | Effect |
| system-architecture | The mockup shows what the user sees. Architecture designs how services deliver it. Together they cover front and back. |
| to-tickets | Each mockup screen/story maps to a ticket. "The edit-report screen" becomes a ticket with the mockup as visual reference. |
| implement-loop | The implement subagent sees the PRD body. If the mockup exists, the visual spec is in .scratch/ for reference. |
to-spec | Produces the PRD text. Product review renders it visually — same source, different medium. |
domain-modeling | If UI vocabulary is ambiguous during mockup (e.g., "what do we call this component?"), call domain-modeling. |
When to skip
| Type | Reason |
| API-only / backend-only | No UI to mock up |
| Bug fix (UI) | The fix is visual but the existing UI is unchanged in layout |
| Mechanical refactor | No UI change |
| Follows existing pattern exactly | Mockup would be identical to existing pages — skip and note it |
| Data-only feature | New field on model, no user-facing change |
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

