myoshi2891

Guide and automate the monthly updates for the 30 application screens and guides. Ensures pricing data, release versions, API specs, external links, and local tests are fully synchronized with minimal diff.

Migrate static HTML guide pages from legacy/ to web-next/ App Router page.tsx using TDD. Phase A–F is fully complete. Use this skill for new guide pages or maintenance of existing migrated pages. TRIGGER when the user says any of the following (Japanese or English): - "新規ガイドページを追加" / "ガイドページを移行" / "ページを保守" - "add new guide page" / "migrate guide page" / "nextjs page migration" Applies project-specific patterns: SiteHeader, DisclaimerBanner, nav-links.ts, CSS Modules, shiki build-time highlighting, Mermaid lazy loading.

Survey any codebase as a senior advisor and produce prioritized, self-contained implementation plans for OTHER models/agents to execute. Strictly read-only on source code — never implements, fixes, or refactors anything itself. Use when asked to audit a codebase, find improvement opportunities (bugs, security, performance, test coverage, tech debt, migrations, DX), suggest features or where to take the project next (roadmap, product direction), or generate handoff plans for another agent to implement.

Comprehensive guide and rules for formatting Markdown files to comply with the project's `.markdownlint.json` configuration. Addresses common errors like MD031, MD012, and MD047. Trigger: Markdownリント, markdownlint, markdown formatting, MD031, MD012, MD047, blanks-around-fences.

Use this skill to fix Mermaid diagram syntax errors, rendering issues, and dark-mode styling problems inside HTML files. Trigger when the user mentions: "mermaid error", "Syntax error in text", "mermaid not rendering", "diagram is broken", "all diagrams crashed", "文字が読めない", "はみ出している", "図が醜い", "ダークモードで見にくい", "シーケンス図が切れている", "マインドマップの色がおかしい", or references a Mermaid version error (e.g. "mermaid version 10.9.5"). Covers: syntax fixes, SVG sizing, dark-mode coloring (mindmap/sequence), foreignObject CSS limitations, and font-loading timing.

Use this skill to apply Subresource Integrity (SRI) integrity hashes and crossorigin attributes to external CDN assets (Prism, Tabler Icons, Mermaid), or to fix diagram rendering issues caused by Mermaid v10+ ESM migration vs v9 UMD API compatibility problems. Trigger when the user mentions: "SRI", "integrity", "crossorigin", "改ざん対策", "セキュリティ強化", "CDN", "Prism", "Tabler Icons", "mermaid", "mermaid 描画", "undefined" in diagrams, or refers to CDN vulnerability/hashes.

Run a full data refresh: scrape pricing data, update exchange rates, and build the frontend. TRIGGER when the user says any of the following (Japanese or English): - "データをフル更新して" / "全体をビルドして" / "最新の価格を取得して" - "full update" / "rebuild all" / "refresh pricing data" Executes update.sh (full scrape + build) or update.sh --no-scrape (exchange rate only).

Sync and verify type definitions between scraper/src/scraper/models.py (Pydantic) and web-next/types/pricing.ts (TypeScript). TRIGGER when the user says any of the following (Japanese or English): - "型を同期して" / "models.py を変更した" / "pricing.ts を更新して" - "スキーマが一致しているか確認して" - "sync types" / "type sync" / "check type parity" Pydantic models.py is the Single Source of Truth; TypeScript side must match.

Keeps spec documents and coverage reports in sync after test implementation, feature implementation, or phase completion. Updates quantitative metrics (test count, suite count, type error count), regenerates coverage-dashboard.html, and syncs functional specs (04-interfaces.md, 05-workflows.md, 08-open-questions.md) and progress trackers (docs/design/*/PROGRESS.md, docs/PROGRESS.md next-actions) when new server actions, pages, or workflows were implemented. Triggered by: "仕様書更新", "ドキュメント更新", "テスト後の仕様同期", "spec sync after test", "update docs after test", "テスト実装後の更新", "カバレッジダッシュボード更新", "update coverage dashboard", "フェーズ完了", "機能実装後の仕様同期", "phase complete", "feature sync".

Runs unit tests (Jest), optional E2E tests (Playwright), TypeScript type check, and ESLint, then reports results with pass/fail details, coverage, and a clear commit-readiness judgment. Recommended before every commit. Triggered by: "テスト実行", "テスト確認", "カバレッジ確認", "品質チェック", "テストチェック", "コミット前確認", "test run", "run tests", "check coverage", "quality check", "pre-commit check".

Regenerates the data-model ER diagram from prisma/schema.prisma and visually declutters edge routing (lines must not overlap entity boxes) without ever hand-editing the generated .drawio. Manual layout intent is captured in the layout-overrides side-car so it survives regeneration. Triggered by: "ERD調整", "ER図調整", "データモデル図", "data-model.drawio", "drawio調整", "図の線が重なる", "線がボックスを突き抜ける", "erd:generate後の調整", "ERD再生成", "モデル追加後の図更新", "adjust ERD diagram", "fix ER diagram routing", "regenerate data model diagram", "erd layout".

Diagnoses CI-only test flakes — tests that pass locally but fail intermittently on GitHub Actions. Triaged via gh CLI log inspection and event-pair comparison (push vs pull_request on the same SHA). Recommends minimal diagnostic instrumentation (Jest --verbose --ci) before any code change. Triggered by: "CI が落ちる", "CI でだけ失敗", "テストがフレーキー", "ローカルでは通る", "間欠的に失敗", "CI flake", "intermittent failure", "flaky test", "passes locally fails in CI".

Generates consistent server action templates in src/queries/ by learning existing patterns. Produces implementation file, Zod schema addition, and unit tests (AAA pattern) in a single scaffold. Only triggered when a new server action is explicitly requested. Triggered by: "新しいサーバーアクション", "サーバーアクション追加", "server action作成", "新しいserver actionを追加", "add server action (query)", "new server action", "add server action", "scaffold action", "create query".

Detects and reports divergences across four layers: (1) specs ↔ implementation, (2) steering rules ↔ rules/skills/agent docs, (3) top-level docs (CLAUDE.md / README.md / GEMINI.md / .agent/) ↔ steering, and (4) SSOT duplication. Never auto-fixes — always reports to the human for judgment. Follows SDD (Spec-Driven Development) principles where specs are the single source of truth, while extending coverage to the operational layer that teaches the rules. Triggered by: "仕様確認", "仕様同期", "spec確認", "仕様書チェック", "仕様と実装の整合性", "仕様乖離", "ルール確認", "規約整合性チェック", "規約浸透チェック", "ドリフト確認", "skill 整合", "rule 整合", "spec check", "sync check", "check spec", "verify spec alignment", "rule sync", "skill sync", "drift check", "仕様乖離をチェック".

世界トップクラスのQAエンジニアとして、Multi-Vendor ECプロジェクトのテスト設計・実装・品質検証を行う。 単なるテスト作成にとどまらず、境界値、異常系、冪等性、競合状態を網羅した「壊れないシステム」を保証する。 「テストを書いて」「品質を上げて」「エッジケースを考慮して」といった指示に適用する。

Backend implementation and fixes for multi-vendor e-commerce projects, including server actions, database and authentication.

Complete workflow for migrating static HTML pages to Next.js App Router page.tsx in the QA_Studies project. Covers CSS variable mapping (HTML vars to Tailwind v4 @theme tokens), page-specific CSS extraction, Header.tsx navigation updates, and CLAUDE.md documentation. Extends the global html-to-nextjs-migration skill with project-specific knowledge including font loading via next/font/google, design token alignment, and accessibility patterns. Trigger: HTMLマイグレーション, ページ移行, HTML変換, 静的HTML移行, CSS変数マッピング, unit-testing-guide.html migration, new page creation from HTML, HTMLからpage.tsx, mat/mbt/tas guide migration.

Use this skill to fix Mermaid diagram syntax errors inside HTML, Markdown, and TSX files. Trigger when the user mentions: "mermaid error", "Syntax error in text", "mermaid not rendering", "diagram is broken", "all diagrams crashed", or references a Mermaid version error (e.g. "mermaid version 10.9.5"). Fixes formatter-induced indentation pollution and statement concatenation that break Mermaid v10 parsing.

Use this skill to fix Mermaid diagram syntax errors inside HTML, Markdown, and TSX files. Trigger when the user mentions: "mermaid error", "Syntax error in text", "mermaid not rendering", "diagram is broken", "all diagrams crashed", or references a Mermaid version error (e.g. "mermaid version 10.9.5"). Fixes formatter-induced indentation pollution and statement concatenation that break Mermaid v10 parsing.

Ensures all project specification documents (CLAUDE.md, GEMINI.md, README.md, docs/MIGRATION_PROGRESS.md, docs/REUSABLE_PROMPTS.md, docs/coverage-dashboard.html, and individual skills/rules in .claude and .gemini) are kept synchronized and updated with the latest project status, test counts, paths, and clear pending migrations. Requires updating the "Last Updated" or "Updated YYYY-MM-DD" timestamp in each document whenever a change is made. Trigger: 仕様書の更新, 更新漏れ確認, ドキュメント同期, 各仕様書を更新, 各仕様書の更新漏れがないか, docs sync, 仕様書同期, spec sync, test追加, テスト追加, 依存関係の更新, CIワークフロー変更, 設定ファイル追加, セッション終了, 再開プロンプト, MIGRATION_PROGRESS, CLAUDE.md, GEMINI.md, README.md, REUSABLE_PROMPTS.md, docs-sync, spec-sync, 最終更新日, Last Updated.

Ensures all project specification documents (CLAUDE.md, GEMINI.md, README.md, docs/MIGRATION_PROGRESS.md, docs/REUSABLE_PROMPTS.md, docs/coverage-dashboard.html, and individual skills/rules in .claude and .gemini) are kept synchronized and updated with the latest project status, test counts, paths, and clear pending migrations. Requires updating the "Last Updated" or "Updated YYYY-MM-DD" timestamp in each document whenever a change is made. Trigger: 仕様書の更新, 更新漏れ確認, ドキュメント同期, 各仕様書を更新, 各仕様書の更新漏れがないか, docs sync, 仕様書同期, spec sync, test追加, テスト追加, 依存関係の更新, CIワークフロー変更, 設定ファイル追加, セッション終了, 再開プロンプト, MIGRATION_PROGRESS, CLAUDE.md, GEMINI.md, README.md, REUSABLE_PROMPTS.md, docs-sync, spec-sync, 最終更新日, Last Updated.

Comprehensive guide and rules for formatting Markdown files to comply with the project's `.markdownlint.json` configuration. Addresses common errors like MD031, MD022, MD032, and MD047. Trigger: Markdownリント, markdownlint, markdown formatting, MD031, MD022, blanks-around-fences, blanks-around-headers, MD047.

Comprehensive guide and rules for formatting Markdown files to comply with the project's `.markdownlint.json` configuration. Addresses common errors like MD031, MD022, MD032, and MD047. Trigger: Markdownリント, markdownlint, markdown formatting, MD031, MD022, blanks-around-fences, blanks-around-headers, MD047.

Complete workflow for migrating static HTML pages to Next.js App Router page.tsx in the QA_Studies project. Covers CSS variable mapping (HTML vars to Tailwind v4 @theme tokens), page-specific CSS extraction, Header.tsx navigation updates, and CLAUDE.md documentation. Extends the global html-to-nextjs-migration skill with project-specific knowledge including font loading via next/font/google, design token alignment, and accessibility patterns.

Keeps spec documents and design logs in sync after test implementation or fixes. Updates quantitative metrics (test count, suite count, type error count). Triggered by: "仕様書更新", "ドキュメント更新", "テスト後の仕様同期", "spec sync after test", "update docs after test", "テスト実装後の更新".

Detects and reports divergences across implementation, design, and specifications. Never auto-fixes — always reports to the human for judgment. Triggered by: "仕様確認", "仕様同期", "spec確認", "仕様書チェック", "仕様と実装の整合性", "仕様乖離", "ルール確認", "規約整合性チェック", "ドリフト確認", "skill 整合", "rule 整合", "spec check", "sync check".

Runs unit tests (Vitest), optional E2E tests (Playwright), TypeScript type check, and ESLint, then reports results with pass/fail details and commit readiness. Triggered by: "テスト実行", "テスト確認", "カバレッジ確認", "品質チェック", "テストチェック", "コミット前確認", "test run", "run tests", "check coverage", "quality check", "pre-commit check".

Diagnoses CI-only test flakes — tests that pass locally but fail intermittently on GitHub Actions. Triaged via gh CLI log inspection. Recommends minimal diagnostic instrumentation (Vitest --reporter=verbose) before any code change. Triggered by: "CI が落ちる", "CI でだけ失敗", "テストがフレーキー", "ローカルでは通る", "間欠的に失敗", "CI flake", "intermittent failure", "flaky test", "passes locally fails in CI".

Analyzes existing production code, identifies untested paths and edge cases, and generates comprehensive test cases using the project's test infrastructure. Supports unit and component (Vitest + RTL), and E2E (Playwright) test generation. Triggered by: "テスト追加", "テストケース追加", "テスト補完", "テスト生成", "カバレッジ改善", "不足テスト", "テストギャップ", "add tests", "generate tests", "write tests for".

Verify and update documentation after a git push containing feat/fix/refactor commits. TRIGGER when the user says any of the following (Japanese or English): - "ドキュメントを同期して" / "ドキュメント更新漏れを確認" - "push したからドキュメント確認して" / "仕様書を最新化" - "docs が古い" / "stale docs" / "doc sync" / "docs sync" - "/check-docs-sync" Detects stale sections in CLAUDE.md, README.md, TEST_PLAN.md, and .claude/skills/*/SKILL.md by comparing git-diff output against actual file state. Updates only the sections that are out of date. Does NOT run tests.

Ensures all project specification documents (CLAUDE.md, README.md, docs/progress.md, and individual skills/rules in .claude) are kept synchronized and updated with the latest project status, test counts, paths, and clear pending tasks. Requires updating the "Last Updated" or "Updated YYYY-MM-DD" timestamp in each document whenever a change is made. Trigger: 仕様書の更新, 更新漏れ確認, ドキュメント同期, 各仕様書を更新, 各仕様書の更新漏れがないか, docs sync, 仕様書同期, spec sync, test追加, テスト追加, 依存関係の更新, CIワークフロー変更, 設定ファイル追加, セッション終了, 再開プロンプト, progress.md, CLAUDE.md, README.md, docs-sync, spec-sync, 最終更新日, Last Updated.

Comprehensive guide and rules for formatting Markdown files to comply with the project's `.markdownlint.json` configuration. Addresses common errors like MD031, MD012, and MD047. Trigger: Markdownリント, markdownlint, markdown formatting, MD031, MD012, MD047, blanks-around-fences.

Run all tests and builds before committing to verify code quality. TRIGGER when the user says any of the following (Japanese or English): - "コミット前チェック" / "テスト実行" / "ビルドとテスト" / "CI 確認" - "全部テストして" / "コミットしていいか確認" - "pre-commit check" / "run tests" / "check before commit" / "verify build" Validates project build (bun run build) and tests (bun run test:all).

Provides best practices, folder structures, and templates for AI Spec-Driven Development (SDD) Markdown files. Use this skill when creating, updating, or reviewing project spec documents (`spec.md`, `requirements.md`, `design.md`, `tasks.md`), agent configuration files (`CLAUDE.md`, `GEMINI.md`, `rules/`), or other Antigravity/Claude Code specific assets.

Applies general implementation best practices, coding conventions, and architectural constraints. Use when writing new code, implementing algorithms, scaffolding new features, or deciding on technical approaches.

Applies accumulated project-specific codebase review rules and best practices. Use when generating, refactoring, or reviewing frontend HTML, rendering logic, or Python generation scripts (like generate_index.py).

Development guidelines for the Hope 3D experience project. Use when modifying Three.js scenes, GSAP animations, CSS styles, or TypeScript components in the Hope project.

Three.js 3D scene development patterns. Use when creating or modifying 3D scenes, cameras, lights, geometries, materials, or animations in Three.js projects.

CSS theming and styling patterns for Hope project. Use when working with theme switching, CSS variables, or styling UI components.