masaya ishihara
28a4c9114b
fix: address PR review feedback for rules-split-manager skill
...
- Add trigger condition to description (when to use)
- Add concrete input/output example for /rules-init
- Remove date from version info
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com >
2026-03-03 10:40:28 +09:00
masaya ishihara
18360bbe94
feat: add rules-split-manager skill for CLAUDE.md split management
...
Manage CLAUDE.md via split rule files in .claude/rules/. Redirects
learning triad (/lessons, /review-learn, /ci-learn) writes to
numbered split files and auto-merges them back into CLAUDE.md.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com >
2026-03-03 10:27:24 +09:00
masaya ishihara
e62ce9ceea
Merge pull request #101 from xtone/feature/add-crash-analyzer-skill
...
feat: add crash-analyze skill for Firebase Crashlytics triage
2026-02-25 11:54:35 +09:00
masaya ishihara
c3dd877a8f
fix: use passive voice in description trigger keywords
...
Change "以下のキーワードで起動:" to "〜で使用される" for
consistent third-person passive voice per review feedback.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com >
2026-02-25 11:45:26 +09:00
masaya ishihara
1aa1829fa4
fix: use third-person trigger description in skill frontmatter
...
Change "以下の場合に使用:" to "以下のキーワードで起動:" per review
feedback to maintain consistent third-person tone.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com >
2026-02-25 11:38:50 +09:00
masaya ishihara
fcbe1d6c3c
fix: use third-person description in skill frontmatter
...
Address review feedback: change description from imperative form
(〜してください) to third-person declarative form (〜するスキル).
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com >
2026-02-25 11:32:10 +09:00
masaya ishihara
fed63791b2
feat: add crash-analyze skill for Firebase Crashlytics triage
...
Add Android-specific skill that analyzes stack traces from Firebase
Crashlytics and provides triage classification (MONITOR/INVESTIGATE/
FIX/DEFEND) with actionable fix suggestions.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com >
2026-02-25 11:25:09 +09:00
Pregum
4be865fecb
Merge pull request #55 from xtone/feature/add-flutter-golden-test-skill
...
feat: Flutter Golden Test スキルの追加
2026-02-17 17:34:21 +09:00
masaya ishihara
a630536ee4
Merge pull request #100 from xtone/feature/add-ci-learner-skill
...
feat: add ci-learner skill for CI failure pattern learning
2026-02-17 10:12:58 +09:00
masaya ishihara
b381457b32
feat: add ci-learner skill for CI failure pattern learning
...
Third pillar of the learning triad (lessons/review-learn/ci-learn).
Analyzes CI/GitHub Actions failure logs and accumulates patterns
into CLAUDE.md to prevent recurring CI failures.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com >
2026-02-17 09:52:06 +09:00
masaya ishihara
5debb90c34
Merge pull request #99 from xtone/fix/video-keyframe-analyzer-skill-discovery
...
fix: video-keyframe-analyzer がセッション再起動で認識されない問題を修正
2026-02-16 12:44:45 +09:00
masaya ishihara
dd6e677b40
fix: quote SKILL.md description to fix YAML parsing failure
...
video-keyframe-analyzerのSKILL.mdフロントマターのdescriptionが
クォートなしで$, %, 括弧等の特殊文字を含んでおり、YAMLパース失敗で
スキルが認識されない問題を修正。他スキルと同様にダブルクォートで囲む。
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com >
2026-02-16 12:33:43 +09:00
masaya ishihara
217f3abe22
Merge pull request #97 from xtone/feature/improve-review-feedback-learner-filtering
...
feat: review-feedback-learner v1.1 - 対応済みフィルタリング導入
2026-02-16 12:27:44 +09:00
Yoichi Toyota
9bde5d6acc
Merge pull request #98 from xtone/fix/update-code-review-skill-from-ci-feedback
...
fix: update code-review skill based on CI execution feedback
2026-02-16 11:05:05 +09:00
toyota
841f9fa5a4
docs: add allowedTools requirements and workflow example for CI
...
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com >
2026-02-16 11:01:21 +09:00
toyota
0b3bdbdc12
fix: correct external skill name to vercel-react-best-practices
...
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com >
2026-02-16 11:01:18 +09:00
masaya ishihara
61340fa414
feat: add addressed-feedback filtering to review-feedback-learner (v1.1)
...
PR #2163の実データ分析で判明した課題(Bot指摘9件が全て未対応でノイズ化)を解決。
「対応した指摘のみ蓄積」フィルタを導入し、有用なナレッジのみを抽出する。
- GraphQL APIでisResolved/isOutdatedを取得
- Bot vs 人間のコメント自動分類
- 対応済み判定ロジック(resolved→outdated+commit→キーワード)
- APPROVEDレビューbodyからの指摘抽出
- 確度別グループ表示
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com >
2026-02-16 10:54:32 +09:00
Sosuke Okubo
0da1868e42
Merge pull request #39 from xtone/feature/add-ios-development-marketplace
...
feat: iOS開発ツールをClaude Codeマーケットプレイスに追加
2026-02-10 19:44:19 +09:00
Sosuke Okubo
7278d52e45
Merge pull request #40 from xtone/feature/add-figma-to-swiftui-skill
...
feat: add implementing-figma-to-swiftui skill for iOS development
2026-02-10 19:30:45 +09:00
Yoichi Toyota
59860e9c47
Merge pull request #95 from xtone/feature/add-code-review-skill
...
feat: add development-workflows plugin with code-review skill
2026-02-10 13:20:58 +09:00
toyota
219d01ddbe
fix: correct typo in authorization-review filename (portgres → postgres)
...
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com >
2026-02-10 13:14:05 +09:00
toyota
ea93b11c23
fix: limit GitHub comment posting to CI environments only
...
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com >
2026-02-10 13:08:03 +09:00
toyota
88652e9736
docs: add GitHub PR review actions reference
...
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com >
2026-02-10 13:08:03 +09:00
toyota
1533e9c994
docs: update README with current plugin list
...
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com >
2026-02-10 13:08:03 +09:00
toyota
e8fd1dd196
feat: add development-workflows plugin to marketplace
...
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com >
2026-02-10 13:07:08 +09:00
Yoichi Toyota
660f7082c7
Merge pull request #87 from xtone/feature/add-code-review-skill
...
feat: add code-review skill for development workflows
2026-02-10 12:52:55 +09:00
toyota
3239875860
feat: integrate authorization references and improve trigger keywords
...
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com >
2026-02-10 12:50:33 +09:00
toyota
8918509080
docs: add authorization review references for code-review skill
...
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com >
2026-02-10 12:50:30 +09:00
Yudai Katsume
b4fe820aef
feat: Step 0にGolden Testの具体例を追加
...
セットアップ前でも完成イメージが把握できるよう、
ウィジェット → テストコード → 生成ファイル → 実行コマンドの
一連の流れを具体例として提示するように変更。
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com >
2026-02-09 12:30:37 +09:00
Yudai Katsume
9acbac9de6
fix: Step 3 Riverpod使用確認の選択肢を初心者向けに改善
...
既存のテストヘルパーが検出された場合に、選択肢のメリット・デメリットが
不明瞭で初心者が判断しづらい問題を修正。
プロジェクト状態に応じて3パターン(A/B/C)に分岐し、
各選択肢にメリット・デメリット・推奨を明記するように変更。
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com >
2026-02-09 12:13:48 +09:00
masaya ishihara
226c95bfe4
feat: add review-feedback-learner skill for PR review knowledge accumulation ( #94 )
...
* feat: add review-feedback-learner skill for PR review knowledge accumulation
PRレビューで受けた指摘をCLAUDE.mdのルールとして蓄積するスキル。
## 概要
- `/review-learn` でPRレビューコメントからルールを抽出
- カテゴリ自動分類(命名規則、アーキテクチャ、エラーハンドリング等)
- レビュアークレジット付きでCLAUDE.mdに追記
- lessons-md-managerと補完関係(自分の気づき+他者の指摘)
## 複数人リポジトリでの価値
- 新メンバー: 過去の指摘が蓄積済み
- レビュアー: 同じ指摘を繰り返す手間が減る
- チーム: 暗黙知が形式知化
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com >
* fix: change "起動" to "使用" in skill description
Address PR review feedback for terminology consistency.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com >
---------
Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com >
2026-02-09 10:52:03 +09:00
masaya ishihara
5b853af29c
Merge pull request #69 from xtone/feature/add-lessons-md-manager-skill
...
feat: add lessons-md-manager skill
2026-02-06 16:05:28 +09:00
masaya ishihara
ca891ef6a0
refactor: move lessons-md-manager to common_development plugin
...
マーケットプレイス配布のため、正しいディレクトリ構造に移動。
- .claude/skills/lessons-md-manager/ → common_development/skills/lessons-md-manager/
- common_development プラグインの marketplace.json, plugin.json を追加
- ルート README.md にプラグイン情報を追加
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com >
2026-02-06 16:04:59 +09:00
masaya ishihara
c1ae76e67e
fix: rename skill from lessons-md-manager to lessons for /lessons command
...
YAML frontmatterのnameがスラッシュコマンド名になるため、
name: lessons-md-manager → name: lessons に変更。
これにより /lessons でスキルが起動する。
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com >
2026-02-06 16:04:13 +09:00
masaya ishihara
664ad30688
feat: add session-end auto-suggestion for /lessons
...
初回 /lessons 実行時にCLAUDE.mdへ Session Rules を追加提案する機能を追加。
これによりClaudeがセッション終了前に自動で /lessons を提案するようになる。
- 追加モードにステップ2「Session Rulesの確認」を追加
- 初期テンプレートに Session Rules セクションを追加
- v1.1としてバージョン更新
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com >
2026-02-06 16:04:13 +09:00
masaya ishihara
a2bbc79c94
feat: add lessons-md-manager skill for CLAUDE.md knowledge accumulation
...
CLAUDE.mdの「Lessons Learned」セクションを管理するPF非依存スキル。
セッション中の修正・失敗・発見を構造的に蓄積し、チーム全体の知見として共有する。
- /lessons: 学びの追記(カテゴリ自動判定、重複チェック付き)
- /lessons review: 既存ルールのレビュー・整理提案
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com >
2026-02-06 16:04:13 +09:00
masaya ishihara
2655f8d058
Merge pull request #42 from xtone/feature/add-video-keyframe-analyzer
...
feat: Add video-keyframe-analyzer skill for cost-efficient video analysis
2026-02-06 16:03:40 +09:00
masaya ishihara
f880c1b764
refactor: move video-keyframe-analyzer to common_development plugin
...
マーケットプレイス配布のため、正しいディレクトリ構造に移動。
## 変更内容
- skills/video-keyframe-analyzer/ → common_development/skills/video-keyframe-analyzer/
- common_development プラグインの marketplace.json, plugin.json を追加
- ルート README.md にプラグイン情報を追加
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com >
2026-02-06 15:59:12 +09:00
masaya ishihara
6c4485a516
Add real-world use case and user feedback improvements
...
Based on interview feedback from DOCOMO AI HOME Issue #303 usage.
## Documentation Updates
- Added real-world use case: Issue #303 card transition flicker bug
- Demonstrates "log-invisible problems" detection capability
- Added troubleshooting for python/python3 command
- Added pip3 install instructions
## Key Insights from Use Case
- Discovered root cause in 30 minutes (vs 2-3 hours with manual debugging)
- Found problem invisible in logs: "ContentRecommendExpandedContent's AI message area"
- Transformed subjective user feedback ("feels wrong") into objective analysis
- 86.7% cost reduction (113 frames → 15 frames)
## User Evaluation
- Usability: 4/5
- Keyframe Precision: 5/5
- Processing Speed: 5/5
- Overall Satisfaction: 5/5
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com >
2026-02-06 15:58:17 +09:00
masaya ishihara
ff5f5e72b3
Address code review feedback
...
Implements improvements based on automated code review comments.
## Security & Validation
- Add input validation for video files (existence, type, extension check)
- Add parameter range validation (threshold: 0-255, quality: 0-100, etc.)
- Support 8 video formats (.mp4, .mov, .avi, .mkv, .webm, .m4v, .flv, .wmv)
## Code Quality
- Add try-finally blocks to all VideoCapture usage (4 functions)
- Prevent resource leaks on exceptions
- Fix hardcoded paths in README (use relative paths)
## Changes
- extract_keyframes.py:
- Added parameter validation functions for argparse
- Added try-finally to: analyze_video_samples, extract_keyframes_distributed,
extract_keyframes_sequential, extract_keyframes_by_timestamps
- Added video file validation in main()
- README.md:
- Changed /Users/m.ishihara/... to $(pwd) and ~/.config/...
Addresses review comments from automated PR review.
All High Priority items completed.
2026-02-06 15:58:17 +09:00
masaya ishihara
456f3899fb
Add --timestamps and --ensure-full-coverage options
...
Implements P1 and P2-A requirements from improvement document.
## --timestamps option
- Extract frames at specific time points (e.g., "0.0,0.5,1.0")
- Useful for animation spec verification
- Allows precise comparison between spec and implementation videos
- Automatically sorts timestamps and validates ranges
## --ensure-full-coverage option
- Guarantees full video coverage by ignoring threshold
- Extracts at least one frame per segment
- Fixed segment calculation bug (last segment now includes all remaining frames)
- Shows time coverage percentage and warns if < 95%
- Useful for subtle animations (fade-in, fade-out, etc.)
## Bug fixes
- Fixed segment calculation in distributed mode
- Last segment now processes all remaining frames instead of dropping them
- Improved from 56% → 78% coverage on test video
## Documentation
- Added comprehensive usage examples for both options
- Added parameter descriptions to README
- Added recommended use cases
Tested with 11.72s lock screen animation video.
2026-02-06 15:58:17 +09:00
masaya ishihara
d233f1f363
Add audio recognition feature with Whisper integration
...
- Integrate OpenAI Whisper for audio transcription
- Add --with-speech and --speech-model options
- Generate markdown report with timeline + full transcription
- Fix timestamp bug (was showing 0.03s-0.67s instead of actual times)
- Fix video duration display bug (was always 0.00s)
- Update README.md with audio recognition documentation
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com >
2026-02-06 15:58:17 +09:00
masaya ishihara
75b80c2fd4
docs: Add 10-minute video length optimization notice
...
10分以内の動画に最適化されている旨をドキュメントに追記
## 変更内容
- SKILL.md: 動画の長さ制限セクションを追加
- 10分以内の動画に最適化されている旨を明記
- 動画の長さ別の推奨設定を追加
- README.md: 同様の情報を追加
- Notion開発スケジュールDB: 概要を更新
## 推奨設定
| 動画の長さ | 推奨max-frames | 処理時間目安 |
|-----------|---------------|-------------|
| 〜10秒 | 15-20 | 数秒 |
| 10秒〜1分 | 20-30 | 20-60秒 |
| 1-5分 | 30-40 | 1-5分 |
| 5-10分 | 40-50 | 5-10分 |
開発用途(画面操作録画、UIテスト、デモ動画など)を想定。
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com >
2026-02-06 15:58:17 +09:00
masaya ishihara
97a8f6d271
feat: Add video-keyframe-analyzer skill for cost-efficient video analysis
...
動画キーフレーム抽出スキルを追加
## 機能
- キーフレーム抽出による93%コスト削減(450円 → 24-30円/5秒)
- 自動モード:動画タイプ判定と最適パラメータ設定
- 4つのプリセット(scroll, ui-transition, high-quality, cost-optimize)
- 均等分割モード:動画全体をカバー
- 閾値自動調整:差分の中央値から最適値を計算
## 実装内容
- extract_keyframes.py: メインスクリプト
- extract_keyframes_distributed(): 均等分割モード
- auto_adjust_threshold(): 閾値自動調整
- detect_video_type(): 動画タイプ自動判定
- SKILL.md: Agent Skills形式スキル定義
- README.md: 使い方とユースケース
- requirements.txt: 依存パッケージ
## テスト結果
- 動画: 2.23秒、88フレーム(39.55fps)
- 抽出: 15フレーム(削減率83%、カバレッジ100%)
- 動画タイプ: scroll(自動判定)
- 閾値: 100(自動調整、デフォルト30から)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com >
2026-02-06 15:58:17 +09:00
Yudai Katsume
63fa294307
refactor: integrate flutter-golden-test with main branch structure
...
- Merge main branch to sync with PR #91 changes
- Add flutter-golden-test to marketplace.json (v0.3.0)
- Update flutter_development/README.md with flutter-golden-test info
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com >
2026-02-06 13:14:03 +09:00
Yudai Katsume
a9f42a7142
Merge remote-tracking branch 'origin/main' into feature/add-flutter-golden-test-skill
2026-02-06 13:09:59 +09:00
Pregum
318e1f644f
Merge pull request #91 from xtone/feature/integrate-screen-spec-generator-to-flutter-development
...
refactor: flutter_screen_spec_generatorをflutter_developmentに統合
2026-02-06 13:05:46 +09:00
Yudai Katsume
12bb390f99
fix: improve SKILL.md description with trigger conditions
...
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com >
2026-02-06 12:55:52 +09:00
Yudai Katsume
8c9a546285
docs: update flutter_development README with new structure
...
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com >
2026-02-06 12:47:36 +09:00
Yudai Katsume
1865265f94
docs: add README.md for flutter-widget-assistant skill
...
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com >
2026-02-06 12:47:32 +09:00