Guide · Agent skill
How to use the critique-theater design review agent skill
Updated · written with AI from the project's own documentation and checked against our catalogue data
critique-theater is an agent skill for Open Design, an agent skill that makes an AI agent review a design or other visual artifact before handing it over. It checks the work across five dimensions, fixes what falls short, and reports the outcome in plain prose. This guide explains what the skill does, how to install it, how to trigger it, and where its limits are, so you can use it on your own layouts and copy.
Design Quality Reviewnexu-io/open-design · Apache-2.0 Download the skill Quick start
- 1Download the ZIP. Get the critique-theater ZIP from this page. It contains open-design.json and SKILL.md.
- 2Unzip into a skills folder. Unzip it into ~/.claude/skills/ for personal use, or into .claude/skills/ inside a project.
- 3Or upload in the Claude apps. If you use the Claude apps, go to Settings → Capabilities → Skills and upload the skill there.
- 4Confirm a design system is active. Brand checks need an active design system. Make sure one is in place before you rely on the brand dimension.
- 5Generate your artifact. Have the agent produce the design, landing page or copy first, since the skill reviews work that already exists.
- 6Ask for the review. Request a quality review in plain language, naming the artifact and any concerns such as accessibility or filler.
- 7Read the report. Check the prose summary of what was checked, fixed and left alone, and review the changes yourself.
What the critique-theater skill makes the agent do
After the agent produces a design or other artifact, the skill asks it to judge that work from five angles: craft, critique, brand, accessibility and copy. Craft covers layout, hierarchy, spacing, alignment and visual rhythm. Critique asks whether the artifact actually solves the stated problem. Brand checks the active design system's tokens, typography and voice. Accessibility covers contrast ratios, touch-target sizes, focus rings, labels on controls and dialogs, keyboard reachability and heading order. Copy looks for filler, placeholder phrasing and marketing padding.
The agent records each concrete defect and fixes it in the same turn. Fixes repeat while defects remain, with three passes described as a sensible ceiling. If something cannot be fixed, the agent states it along with the reason so you can decide.
Install the skill from the ZIP or the Claude apps
Download the ZIP from this page and unzip it. You can place the skill in one of two folders. For a personal setup, use ~/.claude/skills/. For a single project, use the .claude/skills/ folder inside that project. The skill's files are open-design.json and SKILL.md.
If you use the Claude apps instead, upload the skill under Settings → Capabilities → Skills.
- Personal install: ~/.claude/skills/
- Project install: .claude/skills/ inside your project
- Claude apps: Settings → Capabilities → Skills
Check what you need before you start
The skill needs an agent runtime that loads SKILL.md skills. It also refers to an active design system. Brand checks only mean something when there is a design system to check against, and that system is not included in these files. Without one, the brand dimension has nothing concrete to test.
The skill runs no scripts and uses no network access, according to its published details.
- An agent runtime that loads SKILL.md skills
- An active design system for meaningful brand checks (not supplied with this skill)
Trigger the skill with natural requests
The skill is meant to run after the agent has produced an artifact. Ask for the review in ordinary language once the design or copy exists. The skill's own description names the trigger purpose: a five-dimension design quality review that scores the artifact and fixes shortfalls before handing it over.
Requests like these fit the skill's purpose. Replace the placeholders with your own material.
- Review the layout you just made for craft and brand before I share it.
- Check this landing page for accessibility problems like contrast and missing labels, then fix them.
- Look at this campaign copy and cut any filler or placeholder wording.
- Do a quality review of the screen you generated and tell me what you changed and what you left alone.
What you get back
The skill asks for a plain-prose report in the user's language. It says what was checked, what was fixed, and what was deliberately left alone. It does not produce a formal score sheet, a scored panel, a structured review transcript or a review file. No critique.json is written into the project.
Treat the report as the agent's own account of its work. The skill itself notes that the agent reviews its own output and may miss issues.
Safety notes and tips
The skill runs no scripts and uses no network, so it only changes the artifact through the agent's normal editing. Review the changes yourself, because the agent judges contrast and labels by judgment rather than by running automated accessibility tools.
Get better results by supplying your design system and stating the goal clearly. The critique dimension depends on knowing the problem the artifact should solve.
- Provide the design system the brand checks should follow.
- State the audience and goal so the critique dimension has something to test against.
- Read the final prose report and check the items the agent chose to leave alone.
- For compliance sign-off, use a human reviewer or a formal audit instead.
What people use it for
- Designers getting a second pass on a generated layout before sharing it with a client or team.
- Developers building UI who want accessibility and brand checks applied during generation.
- Marketers tightening landing page or campaign copy and visuals for clarity and filler.
Frequently asked questions
How do I install the critique-theater skill?
Download the ZIP from this page and unzip it into ~/.claude/skills/ for personal use, or into .claude/skills/ inside a project. If you use the Claude apps, upload it under Settings → Capabilities → Skills.
What does the critique-theater skill check?
It reviews the artifact across five dimensions: craft, critique, brand, accessibility and copy. It fixes the defects it finds and explains the outcome in plain prose.
Is critique-theater free to use?
The skill is published under the Apache-2.0 licence. Check the licence file for the terms that apply to your use.
Does critique-theater run scripts or use the internet?
According to its published details, it runs no scripts and uses no network access.
Can I use critique-theater for accessibility compliance?
No. The skill checks contrast and labels by the agent's judgment and does not run automated accessibility tools, so it should not be used for compliance sign-off. Use a human reviewer or formal audit for that.
Do I need a design system to use it?
The skill refers to an active design system for brand checks, and that system is not included in the files. Without one, the brand checks have little concrete to test against.
New agent skills, repos and AI tools — every week
The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.
Our verdict
Use critique-theater if you want the agent to give its own generated designs or copy a structured second pass before you share them, especially for accessibility and brand basics. Pick something else if you need an independent human review, a formal audit, or accessibility compliance sign-off, or if your project has no visual or text artifact to check.
Download the skill