機械可読パターンインデックス¶
全パターンの frontmatter メタを1表に集約(scripts/gen_indexes.py で自動生成)。
コーディングエージェントはこの表から候補を絞り、各 .md を読みに行くとよい。
| ID | パターン | domain | forces | driving_variables | forks | status |
|---|---|---|---|---|---|---|
| A1 | Synchronous Edge Agent|同期エッジ | a-execution | [F12] |
[latency_budget] |
`` | stable |
| A2 | Durable Async Agent|耐久非同期セッション | a-execution | [F1, F7, F15, F17] |
[reversibility, latency_budget, accountability] |
`` | stable |
| A3 | Sync Facade over Async Core|非同期コアの同期ファサード | a-execution | [F1, F12, F7] |
[latency_budget] |
`` | stable |
| A4 | Streaming with Progressive Commit|進捗ストリーミング+遅延コミット | a-execution | [F8, F12] |
[latency_budget, failure_cost] |
`` | stable |
| A6 | Adaptive Timeout & Budget-Bounded Retry|適応タイムアウト+予算律速リトライ | a-execution | [F1, F7, F12] |
[latency_budget, cost_sensitivity, failure_cost] |
`` | stable |
| A7 | Deadline & Budget Cascade|期限・予算のカスケード伝播 | a-execution | [F2, F13] |
[request_value, cost_sensitivity] |
[] |
stable |
| B1 | Deterministic Shell, Probabilistic Core|決定論的な殻・確率的な核 | b-orchestration | [F3, F16] |
[failure_cost, task_variability, accountability] |
`` | stable |
| B2 | Workflow Backbone with Agentic Nodes|骨格固定・ノード単位で自律度選定 | b-orchestration | [F3, F10, F13] |
[task_variability, failure_cost] |
`` | stable |
| B3 | Agentic Loop with Budget|予算付き自律ループ | b-orchestration | [F13, F2] |
[task_variability, cost_sensitivity] |
`` | stable |
| B4 | Planner-Executor-Verifier|計画・実行・検証の分離 | b-orchestration | [F2, F4, F13] |
[failure_cost, task_variability] |
`` | stable |
| B5 | Supervisor-Worker|中央オーケストレーション | b-orchestration | [F6, F13, F15] |
[task_variability, cost_sensitivity] |
`` | stable |
| B6 | Critic/Judge & Sampling-Aggregation|独立検証と多数決 | b-orchestration | [F3, F4] |
[request_value, failure_cost] |
`` | stable |
| B7 | Model Router & Adaptive Effort|モデル段階化と適応的努力配分 | b-orchestration | [F2, F12] |
[cost_sensitivity, request_value] |
`` | stable |
| C1 | Tool Gateway / MCP Broker|ツールゲートウェイ・MCP仲介 | c-tools-security | [F8, F14, F16] |
[input_trust, accountability] |
`` | stable |
| C2 | Read-Free / Write-Gated|読取自由・書込ゲート | c-tools-security | [F8] |
[reversibility, failure_cost] |
`` | stable |
| C3 | Dry-run & Commit / Plan-then-Apply|差分提示してから実行 | c-tools-security | [F4, F8] |
[reversibility, failure_cost] |
`` | stable |
| C4 | Idempotent Command Envelope|冪等コマンド包装 | c-tools-security | [F3, F8] |
[reversibility] |
`` | stable |
| C5 | Capability Lease|短命権限チケット | c-tools-security | [F8, F14] |
[input_trust] |
`` | stable |
| C6 | Confused Deputy Defense|代理の混同防御 | c-tools-security | [F14, F5] |
[input_trust] |
`` | stable |
| C7 | Sandboxed Execution|隔離実行 | c-tools-security | [F14] |
[input_trust] |
`` | stable |
| C8 | Saga / Compensation|補償トランザクション | c-tools-security | [F3, F8] |
[reversibility] |
`` | stable |
| D1 | Tiered Memory|階層化メモリ | d-memory-context | [F4, F6] |
[failure_cost] |
`` | stable |
| D2 | Context Budget Allocator|コンテキスト予算配分 | d-memory-context | [F6, F11] |
[cost_sensitivity] |
`` | stable |
| D3 | Memory Write Gate / Quarantine|メモリ書込ゲート | d-memory-context | [F6, F4, F14] |
[input_trust, failure_cost] |
`` | stable |
| D4 | Memory Decay & Versioned Truth|記憶の減衰とバージョン管理 | d-memory-context | [F6] |
[failure_cost] |
[] |
stable |
| D5 | Prompt Registry / Prompt Artifact|プロンプトの成果物化 | d-memory-context | [F9, F16] |
[accountability] |
[] |
stable |
| D6 | Semantic Cache with No-Cache Zones|禁止領域付きキャッシュ | d-memory-context | [F2, F11] |
[cost_sensitivity, failure_cost] |
[] |
stable |
| E1 | Risk-based Human Approval|リスクベース人間承認 | e-safety-hitl | [F4, F8, F17] |
[reversibility, failure_cost] |
`` | stable |
| E2 | Policy-as-Code Guardrail|ポリシーのコード化 | e-safety-hitl | [F14, F16] |
[failure_cost, accountability] |
`` | stable |
| E3 | Input/Output Guardrail Sandwich|入出力ガードレール | e-safety-hitl | [F4, F5, F10, F14] |
[failure_cost, latency_budget] |
`` | stable |
| E4 | Verified Structured Output|検証済み構造化出力 | e-safety-hitl | [F5, F10] |
[failure_cost] |
`` | stable |
| E5 | Autonomy Ladder / Progressive Autonomy|自律性のはしご | e-safety-hitl | [F4, F8, F17] |
[failure_cost, reversibility] |
`` | stable |
| F1 | Short DB Transaction, Long Agent Session|長セッション・短トランザクション | f-data-integrity | [F1, F3] |
[reversibility] |
`` | stable |
| F2 | Event-sourced / Replayable Runs|イベントソーシングとリプレイ | f-data-integrity | [F3, F15, F16] |
[accountability] |
[] |
stable |
| G1 | Tiered (Hot/Cold) Observability|二層観測 | g-observability-ops | [F2, F3, F11, F15, F16] |
[accountability, cost_sensitivity] |
[] |
stable |
| G2 | End-to-End Tracing|全ホップ分散トレース | g-observability-ops | [F1, F15] |
[accountability] |
[] |
stable |
| G3 | Shadow & Canary|影武者とカナリア | g-observability-ops | [F9, F3] |
[accountability] |
[] |
stable |
| G4 | Eval Harness|評価ハーネス | g-observability-ops | [F3, F4, F9] |
[accountability, cost_sensitivity] |
[] |
stable |
| G5 | Circuit Breaker, Graded Degradation & Provider Abstraction|遮断・縮退・抽象化 | g-observability-ops | [F7, F9, F12] |
[provider_trust, cost_sensitivity] |
`` | stable |