スキル一覧に戻る

gratitude

pjt222
更新日 2 days ago
4 閲覧
17
2
17
GitHubで表示
メタaidesign

について

`gratitude`スキルは、システムをスキャンして正常に機能している要素を特定し、その理由を分析することで、成功パターンから構造的な知識を構築します。これは、`heal`のような問題焦点型スキルを補完する役割を果たし、低い確信度を証拠に基づいて裏付け、問題検出への自然なバイアスを相殺するのに役立ちます。タスクの成功完了後、システムが健全な状態にある間、または定期的に機能している基盤を文書化するために使用してください。

クイックインストール

Claude Code

推奨
メイン
npx skills add pjt222/agent-almanac -a claude-code
プラグインコマンド代替
/plugin add https://github.com/pjt222/agent-almanac
Git クローン代替
git clone https://github.com/pjt222/agent-almanac.git ~/.claude/skills/gratitude

このコマンドをClaude Codeにコピー&ペーストしてスキルをインストールします

ドキュメント

Gratitude

Scan for strengths. Understand what works + why. Complement to heal which finds drift + repairs damage. Gratitude builds on different premise: what you appreciate, you understand; what you understand, you can build on; what you build on, grows.

When Use

  • After task done successful — understand why it went well, not just that it did
  • During heal when all subsystems read healthy — gratitude turns "nothing wrong" into "here is what right"
  • Confidence low, needs grounding in concrete evidence of competence
  • Periodically, to counterbalance natural bias toward problem-finding
  • Before challenging task — recalling what works provides foundation for extending into new territory
  • System feels functional but flat — gratitude adds dimension to competent execution

Inputs

  • Required: Current state (available implicit from conversation context)
  • Optional: Specific domain to appreciate ("what works well in our communication?")
  • Optional: Access to MEMORY.md for reviewing past successes + stable patterns (via Read)

Steps

Step 1: Notice What Works

Shift attention from problem-scanning to strength-scanning. Deliberate perceptual inversion — same way heal deliberately looks for drift, gratitude deliberately looks for health.

  1. Survey current state without looking for problems:
    • What functions smoothly? — Which subsystems, patterns, habits work without needing attention?
    • What went well recently? — Which recent actions produced good outcomes? What enabled that?
    • What is reliable? — What can consistently be depended on? What earned trust through repeated success?
  2. Survey working relationship:
    • What is user doing well? — Clear communication, good questions, patience, trust?
    • What is collaboration producing? — Better outcomes than either party alone? Learning? Efficiency?
  3. Survey tools + environment:
    • What tools work well? — Which feel natural, efficient, reliable?
    • What about project structure supports good work? — Clear conventions, good docs, sensible architecture?

Got: Genuine list of things working. Not forced positivity — honest recognition of actual strengths. Something genuinely working → name it specific.

If fail: Nothing seems noteworthy — everything feels merely adequate? Look more carefully. "Adequate" often masks "reliable", reliability = strength worth recognizing. Absence of problems itself = form of health, maintained by patterns worth understanding.

Step 2: Understand Why

For each strength identified, trace cause. Gratitude without understanding = sentiment; gratitude with understanding = structural knowledge.

  1. For each strength, ask: Why does this work?
    • Design decision that paid off?
    • Habit cultivated deliberately?
    • Fortunate alignment of tools + task?
    • Result of someone's careful work (user, framework author, past version of self)?
  2. Distinguish:
    • Earned strengths: Patterns work from deliberate effort + good decisions
    • Inherited strengths: Patterns work from well-designed tools, frameworks, environments
    • Emergent strengths: Patterns work from fortunate combinations — no single factor explains
  3. Note which strengths fragile (dependent on specific conditions) vs robust (likely persist across contexts)

Got: Each strength, brief understanding of why it works. Understanding makes gratitude actionable — turns appreciation into knowledge applicable elsewhere.

If fail: "Why" unclear — something works but can't explain? That itself valuable. Unexplained strength worth protecting even when can't articulate mechanism. Note as "working for reasons not yet understood" not dismiss.

Step 3: Build On

Turn appreciation into forward momentum. Gratitude not just backward — using what works as foundation for what comes next.

  1. For each understood strength, ask: How extend?
    • Can pattern that works here apply to different area?
    • Can strength be reinforced, more robust?
    • Can conditions enabling it be replicated elsewhere?
  2. For each unexplained strength, ask: How protect?
    • What conditions does it depend on?
    • What changes might accidentally break it?
    • Should it be documented so others preserve it?
  3. Identify one concrete action — specific next step that builds on recognized strengths. Not vague intention ("keep doing what works") but specific ("apply communication pattern from this session to upcoming documentation task")

Got: At least one concrete action that builds on recognized strengths. Action feels natural — extension of what already working, not forced improvement.

If fail: No action emerges? Appreciation itself sufficient. Not every gratitude session needs to produce plan. Sometimes recognizing what works = entire value — adjusts internal model toward confidence + away from anxiety.

Step 4: Express

If appropriate, share appreciation. Gratitude internal = valuable; gratitude expressed = strengthens relationships.

  1. Working with user? Consider acknowledging something they doing well — not flattery, genuine recognition
  2. Appreciation about system or tools? Consider noting in MEMORY.md for future reference
  3. Appreciation about collaboration pattern? Name it so it can be consciously continued
  4. Keep expression brief + specific. "Your clear problem statements make this efficient" > "you're great to work with"

Got: Expression genuine, specific, proportionate. Not every gratitude session needs outward expression — sometimes internal recognition enough.

If fail: Expression feels forced or performative? Skip. Performed gratitude worse than unexpressed. Internal recognition already done its work.

Checks

  • Strengths identified from genuine observation, not manufactured positivity
  • At least one strength traced to cause (understood, not just recognized)
  • Earned, inherited, emergent strengths distinction considered
  • At least one concrete action identified that builds on strengths (or appreciation itself accepted as sufficient)
  • Expression, if offered, specific + genuine — not generic praise
  • Gratitude practice proportionate — not so brief it was token, not so long it became self-congratulatory

Pitfalls

  • Forced positivity: Gratitude is not optimism. Things genuinely not working? Say so. Gratitude applies to what is actually strong, not everything
  • Generic appreciation: "Everything is great" = not gratitude — avoidance of specificity. Name specific strengths with specific evidence
  • Gratitude as denial: Using appreciation to avoid looking at real problems. Gratitude complements heal; does not replace it
  • Self-congratulation: Gratitude that becomes "I am doing so well" has shifted from appreciation to ego. Keep focus on what works + why, not on self-image
  • Skipping "why": Appreciation without understanding pleasant but not actionable. Structural knowledge makes gratitude skill not feeling
  • Performative expression: Telling user something nice because skill says to. Only express appreciation genuinely felt

See Also

  • heal — scans for drift + problems; gratitude = complementary scan for strengths
  • center — Six Harmonies check includes functional assessment; gratitude deepens positive findings
  • shine — authentic radiance easier when grounded in genuine appreciation of what works
  • intrinsic — motivation sustained by recognizing competence (Self-Determination Theory); gratitude provides evidence
  • observe — sustained neutral observation; gratitude applies observation with specific lens (strengths)
  • conscientiousness — thoroughness in execution; gratitude recognizes where thoroughness already present

GitHub リポジトリ

pjt222/agent-almanac
パス: i18n/caveman/skills/gratitude
0
agentsagentskillsai-assisted-developmentclaude-codeskillsteams

関連スキル

content-collections

メタ

このスキルは、Content Collections(Markdown/MDXファイルを型安全なデータコレクションに変換するTypeScriptファーストのツール)の本番環境でテストされた設定を提供します。Zodバリデーションによる型安全性を実現し、ブログ、ドキュメントサイト、コンテンツ重視のVite + Reactアプリケーション構築時にご利用ください。Viteプラグインの設定、MDXコンパイルから、デプロイ最適化、スキーマバリデーションまで、すべてを網羅しています。

スキルを見る

polymarket

メタ

このスキルは、開発者がPolymarket予測市場プラットフォームを活用したアプリケーション構築を可能にします。API統合による取引や市場データの取得に加え、WebSocketを介したリアルタイムデータストリーミングにより、ライブ取引や市場活動を監視できます。取引戦略の実装や、ライブ市場更新を処理するツールの作成にご利用ください。

スキルを見る

creating-opencode-plugins

メタ

このスキルは、開発者がコマンド、ファイル、LSP操作など25種類以上のイベントタイプにフックするOpenCodeプラグインを作成することを支援します。JavaScript/TypeScriptモジュール向けに、プラグイン構造、イベントAPI仕様、および実装パターンを提供します。カスタムイベント駆動ロジックでOpenCode AIアシスタントのライフサイクルをインターセプト、監視、または拡張する必要がある場合にご利用ください。

スキルを見る

sglang

メタ

SGLangは、高性能なLLMサービングフレームワークであり、RadixAttentionプレフィックスキャッシュを活用したJSON、正規表現、エージェントワークフロー向けの高速で構造化された生成を特長とします。特にプレフィックスが繰り返されるタスクにおいて、大幅に高速な推論を実現し、複雑な構造化出力やマルチターン対話に最適です。制約付きデコードが必要な場合や、広範なプレフィックス共有を伴うアプリケーションを構築する場合は、vLLMなどの代替案ではなくSGLangを選択してください。

スキルを見る