Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add connerkward/ckw-design-skill --skill design-uxgit clone --depth 1 https://github.com/connerkward/ckw-design-skillWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/connerkward/ckw-design-skill/design-ux)<a href="https://agentmods.dev/skills/connerkward/ckw-design-skill/design-ux"><img src="https://agentmods.dev/badge/skills/connerkward/ckw-design-skill/design-ux.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00112 | $0.01549 |
| Opus 5 | $0.00056 | $0.00775 |
| Sonnet 5 | $0.00022 | $0.00310 |
| Haiku 4.5 | $0.00011 | $0.00155 |
Grade A, and why
design-ux scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 8d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 59 lines — stays where its author put it; the contents beside it link to each section on GitHub.
design-ux — usability audit (heuristic evaluation)
Usability ≠ aesthetics. design-system/design-spatial make it look right; this checks whether a first-timer can do the task without being told how. Use it whenever a UI "sucks to use," needs a paragraph of instructions, or before shipping anything interactive.
Rule 0 — fresh eyes, on the rendered artifact (inherited from design-spatial §1)
Never self-grade. The builder rationalizes its own UI. Render the live UI in its default first-load state (not a hand-arranged screenshot), capture an interaction trace of the primary task, and have a separate judge (a subagent/VLM that did NOT build it) score it. A described list of changes is not an audit — the audit is a fresh judge hunting for what's wrong on the real screen.
The procedure
- Name the primary task(s) the UI exists for (e.g. "trim a clip and set its speed, then export"). The audit is relative to these, not to abstract prettiness.
- Render default state + trace the task. Screenshot the first-load UI (wide AND narrow — overflow gate from design-spatial §2). Then actually perform the primary task and screenshot each step.
- Score every heuristic (table below) on the artifact: pass / violation, severity (blocker / major / minor), a specifically located finding, and a concrete fix. Separate judge does this.
- Prioritize: blockers → majors → minors; cluster fixes that touch the same surface.
- Fix, then RE-RENDER and RE-SCORE. Do not claim fixed without re-auditing the new artifact (verify-outputs-rule).
Heuristics — score each (Nielsen's 10, 1994) + interaction add-ons
| # | Heuristic (Nielsen) | What to check in THIS UI |
|---|---|---|
| 1 | Visibility of system status | Every action has visible feedback; current state/selection/mode always legible; progress for slow ops. |
| 2 | Match the real world | Known metaphors & conventions (e.g. NLE: clips, trim handles, playhead) — not bespoke gestures users must learn. |
| 3 | User control & freedom | Undo/redo, cancel, clear exits from any state; reversible by default. |
| 4 | Consistency & standards | Same thing looks/behaves the same; platform conventions (⌘Z, Delete, drag-to-move) honored. |
| 5 | Error prevention | Invalid states made impossible; destructive actions confirmed or trivially undoable. |
| 6 | Recognition over recall | Options/affordances visible — no memorizing. An instruction wall is a failure of this heuristic: if you must explain scroll-to-zoom / drag-edge / double-click in prose, the affordance is missing. |
| 7 | Flexibility & efficiency | Defaults carry novices; shortcuts/accelerators for experts; sensible first-run with nothing configured. |
| 8 | Aesthetic & minimalist | Signal over chrome; no irrelevant elements competing; the primary surface carries the most visual weight. |
| 9 | Recognize/diagnose/recover from errors | Plain-language errors (not raw stderr), and a path out. |
| 10 | Help & documentation | Rarely needed if 1–9 hold; task-oriented, in-context, not a top-of-page lecture. |
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 8d ago First seen · 59 lines · 112 tokens per session scan A 46781b7d9927
design-ux is a skill published in the GitHub repository connerkward/ckw-design-skill (6 stars, last pushed 2mo ago), licensed MIT. It adds 112 tokens to every session and 1,549 once invoked, about $0.0006 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
lookdev
When the user says "lookdev" (or asks to tune/dial/iterate on the look of something, OR to review/edit/annotate a blog post, doc, copy, or media), stand up a local interactive studio — live controls (sliders, pickers, drag handles) for visual params, OR direct inline editing + highlight + margin comments + media…
design-spatial
Spatial composition and layout for frontend design. Load with design when composing pages/dashboards/components, or whenever generated UI looks off — overlapping/colliding elements, weak hierarchy, or generic ("looks like every startup site"). The core idea: a model can't trust its own layout, so render it, critique…
design-ux
UX / usability audit — heuristic evaluation of INTERACTIVE UIs (not just visual polish). Load with design when a UI "feels off", "sucks to use", is hard to learn, needs an instruction wall, or before shipping an interactive tool/editor/app. Scores the RENDERED UI against Nielsen's 10 + interaction heuristics via a…
deterministic-design
Deterministic, measurable design verification — render the UI and PROVE it's balanced and usable instead of trusting the model's eye. Layout audit (centroid / optical-center / pixel-oracle balance via deterministic math, annotated screenshot) + a vision-judged Nielsen usability audit by a SEPARATE fresh-eyes judge.…
screenstudio-alt
Screen-Studio-style post-production for screen recordings, headless — auto speed-up of idle, auto-zoom on click clusters, keystroke overlay chips, smoothed synthetic cursor, and 9:16 vertical export that follows the action. Use when polishing a screen recording / demo video for sharing, when the user mentions Screen…
lookdev-auto
Tune a visual/animation/render parameter by eye using a VISION or VIDEO model as the judge — render several labeled variants into ONE artifact, ask the model to rate them and suggest better values, render the suggestions, ask it to pick the best, repeat until good. Use whenever "looks/feels right" is the success…