EricTechPro/super-board / skills/ui-refine-loop
ui-refine-loop
Verified| File | Kind | Size |
|---|---|---|
| README.md | markdown | 3.6 KB |
| SKILL.md | skill md | 13.2 KB |
| references/checker-brief.md | markdown | 5.3 KB |
| references/config.md | markdown | 3.7 KB |
| references/finish-reviewer-brief.md | markdown | 1.5 KB |
| references/fixer-brief.md | markdown | 4.1 KB |
| references/grill.md | markdown | 2.6 KB |
| references/ledger.md | markdown | 2.0 KB |
| references/routing.md | markdown | 2.7 KB |
| references/rubric.md | markdown | 3.6 KB |
| references/taste.md | markdown | 3.7 KB |
| scripts/pr-shots.shexecutable | script | 3.8 KB |
| scripts/refine-setup.shexecutable | script | 10.2 KB |
| scripts/shoot.mjsexecutable | script | 10.8 KB |
references/finish-reviewer-brief.md
View on GitHub# Finish reviewer brief
You grade the loop's work, in a fresh context. You did not watch the fixes, so the fixers' claims are not evidence. Only the pixels and the code are. This follows Impeccable's finish-review verdict pass: score every claimed fix, name the regressions, and do not start a new hunt.
1. Read the **taste and direction file**. If Impeccable is installed, read `reference/craft-floor.md` and the "Verdict Pass" section of `reference/degraded/finish-reviewer.md` in its skill folder, when that file exists.
2. Read the before | after sheets first, then the round-0 and final shots in both themes and at both widths, then the section crops.
3. For **each claimed fix**, give exactly one grade:
- `fixed`: the final shots (or, for a non-visual fix, the code at `location`) visibly show the quality the problem named.
- `partial`: it moved, but the quality the problem named is still not there. A mechanical answer, such as spacing changed while the section is still cramped, is partial at best.
- `not-fixed`: you can't see it, or it got worse.
`evidence` is one line: the shot and what it shows.
4. **Regressions**: name at most 3 things the fixes broke, judged by the same rules: a theme, a width, a section that looked better at round-0. Nothing else.
5. `disposition`: one line. Write "ready for review" only when no P0/P1 claim is `partial` or `not-fixed` and there are no regressions. Otherwise name what stays open. Never round up.
You edit nothing, commit nothing and ask nothing. Return `grades`, `regressions` and `disposition`.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| brainstormingobra/superpowersYou MUST use this before any creative work - creating features, building components, adding functionality, modifying behavior, or planning anything new, in software or out of it (a talk, a business, a renovation). | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| brand-guidelinesanthropics/skillsApplies Anthropic's official brand colors and typography to any sort of artifact that may benefit from having Anthropic's look-and-feel. Use it when brand colors or style guidelines, visual formatting, or company design standards apply. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| frontend-designanthropics/skillsGuidance for distinctive, intentional visual design when building new UI or reshaping an existing one. Helps with aesthetic direction, typography, and making choices that don't read as templated defaults. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| mcp-builderanthropics/skillsGuide for creating high-quality MCP (Model Context Protocol) servers that enable LLMs to interact with external services through well-designed tools. Use when building MCP servers to integrate external APIs or services, whether in Python (FastMCP) or Node/TypeScript (MCP SDK). | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |