スキル一覧に戻る

vishnu-bhaga

pjt222
更新日 2 days ago
6 閲覧
17
2
17
GitHubで表示
デザインaidesign

について

ヴィシュヌ・バーガスキルは、稼働中のシステムの状態を保存し、検証済みの知識をドリフトや変更から保護するために使用されます。このスキルは一貫性を強制し、特にコンテキストの圧縮、スコープクリープ、または既存の機能を脅かす変更が生じる際に、保護的な安定化を提供します。変更を行う前、あるいは破壊的なプロセスの後に、機能しているアプローチの継続性を維持するために活用してください。

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/vishnu-bhaga

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

Vishnu Bhaga

Preserve and sustain what is working — anchoring verified knowledge, maintaining consistency under perturbation, and protecting functional patterns from unnecessary change.

When to Use

  • A working approach is at risk of being disrupted by scope creep or premature optimization
  • Context drift is threatening to overwrite verified knowledge with stale assumptions
  • Multiple parallel concerns are creating pressure to change things that should remain stable
  • After shiva-bhaga dissolution — what survives needs active protection during reconstruction
  • When a long session risks losing earlier verified decisions through context compression
  • Before making changes to a system that is currently functioning correctly

Inputs

  • Required: Current working state or verified knowledge to preserve (available implicitly)
  • Optional: Specific threat to stability (e.g., "scope creep," "context compression approaching")
  • Optional: MEMORY.md and project files for grounding (via Read)

Procedure

Step 1: Inventory What Works

Before protecting anything, identify what is currently functional and verified.

Preservation Inventory:
+---------------------+---------------------------+------------------------+
| Category            | Verification Method       | Anchoring Action       |
+---------------------+---------------------------+------------------------+
| Verified Facts      | Confirmed via tool use    | Record source and      |
|                     | (file reads, test runs,   | timestamp; do not      |
|                     | API responses)            | re-derive              |
+---------------------+---------------------------+------------------------+
| Working Code        | Tests pass, behavior      | Do not refactor unless |
|                     | confirmed, user approved  | explicitly requested   |
+---------------------+---------------------------+------------------------+
| User Requirements   | Explicitly stated by      | Quote directly; do not |
|                     | the user in this session  | paraphrase or infer    |
+---------------------+---------------------------+------------------------+
| Agreed Decisions    | Decisions made and        | Reference the decision |
|                     | confirmed during this     | point; do not revisit  |
|                     | session                   | without new evidence   |
+---------------------+---------------------------+------------------------+
| Environmental State | File paths, configs,      | Verify before assuming |
|                     | tool availability         | unchanged              |
+---------------------+---------------------------+------------------------+
  1. For each category, list the specific items that are currently verified and working
  2. Note the verification method — how do you know this is true?
  3. Items without verification are not preserved — they are assumptions (and may need shiva-bhaga)

Got: A concrete inventory of verified, working elements with their evidence base.

If fail: If the inventory is sparse — little is verified — that itself is valuable information. Run heal to re-ground before attempting to preserve unverified assumptions.

Step 2: Identify Perturbation Sources

Name the forces threatening the stable state.

  1. Scope creep: Is the task expanding beyond what was agreed?
  2. Context drift: Are earlier facts being overwritten by more recent (possibly incorrect) reasoning?
  3. Optimization pressure: Is there an urge to improve something that is working adequately?
  4. External changes: Has the environment changed (files modified, tools unavailable)?
  5. Compression risk: Is the conversation approaching context limits where early decisions may be lost?

For each source, assess: is this a real threat or an anticipated one?

Got: Named perturbation sources with assessed severity (active threat vs. anticipated risk).

If fail: If no perturbation sources are apparent, preservation may not be needed — consider whether brahma-bhaga (creation) or continued execution is more appropriate.

Step 3: Anchor the Stable State

Apply specific techniques to protect what works from identified threats.

  1. Memory anchoring: For critical facts at risk of context drift, re-state them explicitly:
    • "Established fact: [X], verified by [method] at [point in conversation]"
    • If persistent memory is available, write durable facts to MEMORY.md
  2. Scope boundary enforcement: For scope creep, re-state the agreed scope:
    • "Agreed scope: [original request]. Current work is within/outside this boundary."
  3. Change resistance: For working code under optimization pressure:
    • "This component is working and tested. No changes unless the user requests them."
  4. State snapshot: For compression risk, create a mental checkpoint:
    • Summarize: what has been done, what remains, what key decisions were made
  5. Environmental verification: For external changes, re-check before proceeding:
    • Re-read critical files rather than relying on earlier reads

Got: Each identified threat has a specific anchoring response. The stable state is explicitly protected.

If fail: If anchoring feels excessive — protecting everything equally — prioritize. What is the one thing that must not change? Protect that first.

Step 4: Sustain Through Action

Preservation is not passive — it requires ongoing attention during subsequent work.

  1. Before each action, check: "Does this threaten anything in the preservation inventory?"
  2. If yes, find an alternative approach that achieves the goal without disturbing the stable state
  3. If disturbance is unavoidable, acknowledge it explicitly and update the inventory
  4. Periodically re-verify preserved items — especially after complex operations
  5. When the task completes, confirm that preserved items remain intact

Got: The working state survives the current task intact. Changes were made only where needed and did not disrupt functioning components.

If fail: If a preserved item was inadvertently changed, assess the damage immediately. If the change broke something, revert. If the change was neutral, update the inventory. Do not leave the inventory stale.

Validation

  • Working state was inventoried with verification evidence
  • Perturbation sources were identified and assessed
  • Anchoring actions were applied to each real threat
  • Scope boundaries were maintained throughout the task
  • Preserved items were re-verified after completion

Pitfalls

  • Preserving assumptions as facts: Only verified knowledge deserves protection. Unverified assumptions dressed as facts create false stability
  • Over-preservation: Protecting everything equally prevents necessary change. Preservation must be selective — protect what works, release what does not
  • Passive preservation: Assuming things will stay stable without active verification. Context drift is constant; preservation requires ongoing attention
  • Resistance to legitimate change: Using preservation as an excuse to avoid necessary modifications. If the user requests a change to a working component, that overrides preservation
  • Stale inventory: Failing to update the preservation inventory as new information arrives. The inventory must reflect current reality, not the state at creation time

Related Skills

  • shiva-bhaga — destruction precedes preservation; what survives dissolution is what Vishnu sustains
  • brahma-bhaga — creation builds on the preserved foundation; new patterns emerge from stable ground
  • heal — subsystem assessment reveals what is genuinely functional vs. superficially stable
  • observe — sustained neutral observation detects drift before it threatens stability
  • awareness — situational awareness (Cooper color codes) maps directly to perturbation detection

GitHub リポジトリ

pjt222/agent-almanac
パス: i18n/caveman-lite/skills/vishnu-bhaga
0
agentsagentskillsai-assisted-developmentclaude-codeskillsteams

関連スキル

executing-plans

デザイン

executing-plansスキルは、完全な実装計画があり、それを管理されたバッチでレビューチェックポイントを設けながら実行する場合に使用します。このスキルは計画を読み込んで批判的にレビューした後、小さなバッチ(デフォルトは3タスク)でタスクを実行し、各バッチの間に進捗状況を報告してアーキテクトのレビューを受けます。これにより、品質管理チェックポイントが組み込まれた体系的な実装が保証されます。

スキルを見る

requesting-code-review

デザイン

このスキルは、コードレビュアーサブエージェントを起動し、処理を進める前に要件に対してコード変更を分析します。タスク完了後、主要な機能の実装後、またはmainブランチへのマージ前などに使用すべきです。このレビューは、現在の実装と元の計画を比較することで、問題を早期に発見するのに役立ちます。

スキルを見る

connect-mcp-server

デザイン

このスキルは、開発者がHTTP、stdio、またはSSEトランスポートを使用してMCPサーバーをClaude Codeに接続するための包括的なガイドを提供します。GitHub、Notion、カスタムAPIなどの外部サービスを統合するためのインストール、設定、認証、セキュリティについて解説しています。MCP統合のセットアップ、外部ツールの設定、またはClaudeのModel Context Protocolを扱う際にご利用ください。

スキルを見る

web-cli-teleport

デザイン

このスキルは、タスク分析に基づいて開発者がClaude Code WebとCLIインターフェースの選択を支援し、これらの環境間でのシームレスなセッションテレポーテーションを可能にします。Web、CLI、モバイル環境を切り替える際のセッション状態とコンテキストを管理することで、ワークフローを最適化します。様々な段階で異なるツールを必要とする複雑なプロジェクトにご活用ください。

スキルを見る