skills/design/SKILL.md
Artifact-backed interface critique and polish for hierarchy, typography, layout, density, IA, interaction feel, content, brand fit, and taste. Requires screenshot, URL, rendered artifact, or explicit file plus intent. Use when: "make this look better", "improve the design", "polish the UI", "critique this screen", "design pass", "art direction", "scaffold design". Trigger: /design.
npx skillsauth add phrazzld/spellbook designInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
Critique and improve a rendered artifact against its intent. The core contract is evidence, not generic advice:
Refuse to make a final design judgment from code alone when a rendered surface can be inspected. If rendering is impossible, mark the design unverified.
Delegation floor applies: probe the roster first; dispatch two or more
providers for substantive work; direct solo only for mechanical, emergency,
user-forbidden, or fewer-than-two-providers cases. See
harnesses/shared/AGENTS.md (Roster).
Local lane guidance: Use one lane for proposed direction or implementation and another for cold review of substantive redesign, external-facing polish, or final critique of visible UI changes.
| Intent | Action |
|---|---|
| /design audit or /design critique | Read intent, inspect artifact, return ranked findings. |
| /design polish | Apply the smallest coherent improvement set, then verify render. |
| /design redesign | Propose 2-3 structurally different directions, get convergence, then implement. |
| /design scaffold | Read references/scaffold.md and generate project-local design guidance, DESIGN.md, and design-contract.md only when recurring UI work earns them. |
Use /a11y for WCAG compliance, /qa for behavior verification, /demo for
evidence packaging, and /shape when the product direction itself is unsettled.
For recurring UI work, use the references:
references/scaffold.md for repo-owned DESIGN.md and
design-contract.md provenance scaffolds.references/design-system.md for token and component-system judgment.references/taste-layer.md for aesthetic direction and anti-generic critique.references/anti-slop.md for the checkable list of AI design tells, the
VARIANCE/MOTION/DENSITY dials, and the pre-emit gate.references/interface-polish.md for micro-polish checks.references/ui-surface-routing.md for workflow composition.Lead with the highest-leverage issues. Avoid a laundry list.
## Design Critique
- Intent:
- Artifact inspected:
- Primary issue:
- Recommended direction:
- Specific changes:
- Verification needed:
Each finding names evidence from the artifact and one concrete change. If the issue is only preference, say so; if it blocks comprehension or trust, say that.
Directions must differ structurally, not just by palette:
For each direction, name what it sacrifices. Recommend one.
See harnesses/shared/AGENTS.md (Completion Evidence) for the shared evidence
core; this phase keeps design-specific local fields.
## Completion Gate
- Direction chosen: critique, polish, redesign, or scaffold decision applied.
- Evidence that proves it: screenshot, render, artifact, or visual diff inspected.
- Exact command/path/route exercised: URL, screenshot path, render command, or artifact path inspected.
- Hierarchy/content changes: specific hierarchy or content issue changed or recommended.
- Typography/layout changes: specific type, spacing, density, or layout issue changed or recommended.
- Copy provenance: visible copy inspected for product truth versus agent process leakage.
- Residual risk: remaining design, a11y, or QA risk after inspection.
development
Lightweight evidence-backed retro and catch-up reports for a current repo, branch, PR, backlog slice, or recent agent session. Use when the user asks for a debrief, catch me up, what changed, why it matters, product implications, end-user implications, developer experience implications, current app state, backlog state, workspace state, alternatives considered, or context rebuild after losing the thread. Trigger: /debrief.
testing
Capture agent-session work records as local JSONL audit evidence. Links a backlog/spec, branch, commits, review verdicts, QA/demo evidence, transcript refs, and shipped ref without storing raw private transcripts. Use when: "trace this work", "write work record", "agent session trace", "journal this delivery", "link transcript evidence". Trigger: /trace, /journal.
data-ai
Turn proven agent-session patterns into first-party Harness Kit skills. Use when: "skillify this conversation", "make this into a skill", "generate a skill from current transcript", "extract reusable workflow". Trigger: /skillify.
testing
Run one targeted, read-only architecture or quality critique through a named lens from the shared rubric. Use when: "critique this module", "run an Ousterhout pass", "lens critique", "architecture critique". Trigger: /critique.