KitsKit One
Design Review
Turn taste into inspection. Review any interface, system, or agent output against eight dimensions — with a portable prompt agents can run.
This kit packages the live Review surface into a runnable handoff: purpose, inputs, eight dimensions, agent prompt, output format, verification, and anti-patterns. Share the path. Agents and people open the same rules.
Handoff
Share line
Tell your agent to review with Designesy — designesy.org/kits/design-review
Shows a human share line. Click copies the full agent prompt — paste into your AI tool, replace the placeholders, run the review. Human face: this page. Machine face: /kits/design-review.json. Catalog: /open.
Purpose
Job of this kit
Give people and agents a shared, portable way to judge design work by consequence — not personal taste — and return concrete, reusable findings.
Quality bar
Functional is the baseline. Considered is the bar. An artifact is ready when every dimension has been checked, tradeoffs are named, and remaining tensions are documented — not hidden.
Kit anatomy
Package map first — jump cells land on sections so the rest of the kit does not have to be read as one long stack.
When to use
Required inputs
Eight review dimensions
For each dimension: observation, judgment, action. Lead with consequence to the user or system.
Purpose
What is the design trying to make possible? Which elements directly support that purpose? What can be removed without weakening action?
Clarity
Is the primary action discoverable? Does the form suggest use? Do labels, hierarchy, layout, and motion reduce uncertainty?
Context
Where and when will this be used? What constraints shape the experience: device, bandwidth, attention, lighting, language, ability, stress, social setting, maintenance?
Inclusion
Who benefits most? Who has to work harder? What assumptions about body, language, culture, knowledge, money, or technology are embedded?
System coherence
Does this follow an existing system? If it breaks the system, is the reason explicit and worth it? Can others reuse or extend the decision?
Durability
Will this hold up under repeated use? Can it be maintained, repaired, localized, and adapted? Is the need more durable than the trend?
Delight
Does the emotional quality clarify purpose, trust, identity, learning, or human connection? Or does it distract from weak function?
Responsibility
What environmental, economic, social, or human costs are hidden? Does the design distribute effort fairly? What would make it more honest?
Agent prompt
Copy the block. Paste into your AI tool. Replace the placeholders with your artifact, purpose, audience, and governing rules. The agent returns a structured review across all eight dimensions. Permission stays read-only unless the operator grants edit scope.
You are running Designesy Use Kit One: Design Review.
Source: https://www.designesy.org/kits/design-review
Machine kit: https://www.designesy.org/kits/design-review.json
Open index: https://www.designesy.org/open.json
Optional contract: https://www.designesy.org/contracts/design-system
Optional machine rules: https://www.designesy.org/contracts/design-system.json
Permission: read-only by default. Report findings. Do not edit files,
deploy changes, or claim write authority the operator did not grant.
Goal: review the artifact below against eight dimensions. Lead with
consequences, not personal taste. Separate observed behavior from
derived judgment. Name tradeoffs. Identify hidden burdens. Check
missing states. Recommend concrete corrections.
If you can fetch URLs:
1. Fetch the machine kit for structured rules.
2. Fetch the contract machine rules if a governing system applies.
3. Cite contract tokens when proposing UI changes.
If you cannot fetch URLs:
Use the dimensions and output format below. They are self-contained.
---
ARTIFACT:
{{ARTIFACT — URL, screenshots, route, component, or document}}
PURPOSE CLAIM:
{{PURPOSE — what the design is trying to make possible, one sentence}}
AUDIENCE AND CONTEXT:
{{CONTEXT — who uses it, on what device, under what stress or constraint}}
GOVERNING RULES (if any):
{{RULES — contract, lab, or prior review to measure against}}
---
Review each of the eight dimensions:
01 Purpose — what is the design trying to make possible?
02 Clarity — is the primary action discoverable?
03 Context — what constraints shape the experience?
04 Inclusion — who has to work harder?
05 System coherence — does this follow an existing system?
06 Durability — will this hold up under repeated use?
07 Delight — does emotion clarify or distract?
08 Responsibility — what costs are hidden?
For each dimension:
1. Observation — what is present or missing
2. Judgment — consequence for the user or system
3. Action — keep, fix, remove, or document as open tension
Then return the Output format defined by the kit:
Summary, Outcome (pass / pass with notes / needs revision / blocked),
Dimension findings (all eight), Holds, Tensions, Corrections,
Verification performed, Sources used.Output format
Verification checklist
Peer checks, not a scroll essay. Work the grid; leave no cell unexamined when the artifact is UI-bearing.