Skip to main content
skill CX Codex

pattern-library-curator — Claude Code skill

pattern-library-curator is a Claude Code skill that gathers accepted interface solutions into a project pattern library. It checks recurring decisions for evidence before they are treated as reusable patterns.

MIT ★ 529 updated 2026-11-17

Install into Claude Code · Skills CLI (Vercel) All methods and platforms ↓
npx skills add ShuaiMXu/happyhands --skill pattern-library-curator -a claude-code
Open on GitHub

What pattern-library-curator does

The skill reviews accepted work and returns structured evidence, findings, proposed changes, and limitations. Its Foundation candidate scan checks run chains, source snapshots, QA results, and image hashes. A candidate must involve multiple referenced child components and multiple consumers in the same route, device, and action context. Repeated runs and device copies do not count as extra uses.

Candidates receive needs-evidence; the scan does not add them to Foundation.patterns. Review requires accepted runs, component and token references, and documented exceptions. A one-off solution does not become a pattern by default.

Workbench can draft guidance for one candidate through a separate model adapter and a read-only Codex session. The draft must cite every supplied image and source reference. It is saved as proposal-only and leaves approved Foundation unchanged; professional review still determines the pattern's role and usage rules.

What it helps with

  • turn accepted interface decisions into a project pattern library
  • check evidence before reusing a component
  • document pattern usage, exceptions, and dependencies
  • draft guidance for a pattern candidate

Example prompts for the agent

  • Collect repeated accepted interface decisions into a pattern library for this project.
  • Check the evidence for reusing these components and list the limitations.
  • Draft guidance for this candidate, including when to use it and its exceptions.

How to install pattern-library-curator

1Skills CLI (Vercel)

npx skills add ShuaiMXu/happyhands --skill pattern-library-curator -a claude-code

2Manually: put the files in place

~/.claude/skills/pattern-library-curator/SKILL.md

Copy the whole skill folder; for one project use .claude/skills/pattern-library-curator/

Step by step, with commands for each platform: How to install skills in Claude Code: five ways

Questions about pattern-library-curator

How do I install pattern-library-curator in Claude Code?

Run npx skills add ShuaiMXu/happyhands --skill pattern-library-curator -a claude-code. The skill file is installed at ~/.claude/skills/pattern-library-curator/SKILL.md.

What does pattern-library-curator do?

It finds recurring accepted interface solutions and gathers evidence for a project pattern library. Candidates are not added to the approved library automatically.

What evidence does a pattern need?

The review requires accepted runs, component and token references, and documented exceptions. Professional review must determine the pattern's role and usage rules.

Source text

SKILL.md · MIT
---
name: pattern-library-curator
description: "Curate accepted interface patterns into a governed project library with usage guidance and exceptions."
---

# Pattern Library Curator

Turn repeated accepted solutions into reusable project patterns.

## Current bounded implementation (0.3.0)

The automatic Foundation pack collects candidate compositions from the explicit parent-run chain (maximum 16). It checks canonical repository identity, current complete source/input snapshots, active acceptance, durable passing QA and original image hashes. A candidate needs multiple statically referenced child components, multiple static consumers and distinct declared root identities in the same route/device/action context. Device copies, selector aliases and repeated runs do not multiply uses. Older captures lacking structural root identity cannot be backfilled by inference.

The result remains `needs-evidence`: observed candidates are not inserted into `Foundation.patterns`. Professional review must still establish pattern role, when to use/avoid, exceptions and applicable states/variants. Historical image paths/hashes identify local evidence, not bundled originals; revalidate source runs before model curation. Missing accepted history is an evidence gap, not proof the product has no reusable patterns. See `docs/accepted-pattern-candidates-2026-10-18.md` for bounds and remaining work.

The local Workbench can optionally draft usage guidance from a single candidate. Its separate model adapter revalidates accepted history and current sources, delivers original PNGs plus bounded source context to a read-only Codex session, and validates a dedicated structured-output schema. All supplied image and source references must be cited. `propose` requires usage, avoidance, exception and limitation text; unresolved/non-reusable verdicts cannot supply usage rules. The saved record is always `proposal-only`, with frozen original images and input hashes. It does not replace this deterministic Skill result, modify approved Foundation, complete QA or record human acceptance. Governed pattern admission and professional case calibration remain open. See `docs/pattern-guidance-review-2026-10-18.md`.

## Required evidence

- accepted runs
- component and token references
- documented exceptions

## Decisions

- Require accepted evidence
- Describe when to use and avoid the pattern
- Link dependencies and examples

## Boundary

A one-off solution is not automatically a reusable pattern.

Return structured evidence, findings, proposed updates and limitations. Use `needs-evidence` when required inputs or professional review are missing. Never invent an Experience Score or promote a project observation into a universal rule.

© ShuaiMXu · MIT · ShuaiMXu/happyhands

Similar skills

All skills for Design →

From the same set34

The whole set →
component-inventory component-inventory is a Claude Code skill that maps components to source paths, variants, composition, and known consumers. It supports reuse decisions and drift analysis. CX Codex +1 ★ 529repo stars visual-qa-evaluator Visual-qa-evaluator is a Claude Code skill that prepares evidence for visual QA and regression analysis. It supports matching interface states to evaluate design changes. CX Codex +1 ★ 529repo stars brand-foundation-extractor brand-foundation-extractor is a Claude Code skill that gathers candidate brand principles, voice cues, and visual assets from project files and rendered content. It keeps each finding tied to its source. CX Codex +1 ★ 529repo stars token-contrast-auditor The token-contrast-auditor skill for Claude Code verifies color pair contrast against WCAG standards. It runs during project analysis to check design tokens and CSS styles. CX Codex +1 ★ 529repo stars design-system-drift-detector The design-system-drift-detector skill for Claude Code identifies discrepancies between implementation and design tokens. It detects ungoverned literals, semantic conflicts, and off-scale values. CX Codex +1 ★ 529repo stars fix-planner fix-planner is a Claude Code skill that turns a confirmed finding into a bounded change plan. It ties proposed fixes to project evidence and approved design decisions. CX Codex +1 ★ 529repo stars component-state-auditor The component-state-auditor skill for Claude Code compares documented component states against a versioned rubric in rubric.json. It identifies missing states during UI development and review. CX Codex +1 ★ 529repo stars visual-hierarchy-reviewer visual-hierarchy-reviewer is a Claude Code skill that reviews task hierarchy and design language from screenshots. It runs on interface reviews, including requests to reduce an AI-made look while keeping the product’s brand. CX Codex +1 ★ 529repo stars

Project sponsors

Companies that keep this analytics open