Durable Execution Assistant | 持久執行助手

SkillMonitoring & ops

[UDS] Guide fault-tolerant workflow design with checkpoints, retry policies, and rollback plans. Use when: a long-running workflow keeps failing partway, designing checkpoint granularity, choosing a retry or backoff strategy. Not for: responding to a production incident already in progress — use /incident; deployment rollback mechanics — use /deploy. Keywords: durable execution, checkpoint, retry, backoff, idempotency, rollback, 持久執行, 檢查點, 重試策略.

Available today. Use it from your connected AI after setup.

Connect ahel once, and every AI you use reads what you have installed.

Then ask your AI: use the Durable Execution Assistant | 持久執行助手 skill

What this skill tells your AI

The instructions your AI receives, as published by asiaostrich/universal-dev-standards in skills/durable-execution-assistant/SKILL.md and read by ahel’s review.

Guide fault-tolerant workflow design with automatic recovery, checkpoints, and retry strategies.

引導容錯工作流程設計,包含自動恢復、檢查點與重試策略。

Usage | 使用方式

CommandPurpose用途
/durableStart interactive failure recovery guide啟動互動式故障恢復引導
/durable --checkpointDesign checkpoint strategy設計檢查點策略
/durable --retryConfigure retry policies配置重試策略
/durable --rollbackPlan rollback procedures規劃回滾程序
/durable <workflow>Analyze specific workflow for durability分析特定工作流的持久性

Core Concepts | 核心概念

ConceptDefinition定義
CheckpointSave execution state at known-good points在已知正確的位置儲存執行狀態
RetryAutomatically re-attempt failed operations自動重新嘗試失敗的操作
RollbackRevert to last known-good state on failure失敗時回復到上一個已知正確狀態
IdempotencyOperations produce same result on re-execution操作重新執行時產生相同結果
Circuit BreakerStop retries when failure rate exceeds threshold失敗率超過閾值時停止重試

Failure Recovery Decision Tree | 故障恢復決策樹

Failure detected
├── Is it transient? (network, timeout)
│   ├── Yes → Retry with backoff
│   └── No → Is state corrupted?
│       ├── Yes → Rollback to checkpoint
│       └── No → Is it a logic error?
│           ├── Yes → Stop, report, fix code
│           └── No → Escalate with diagnostics

Retry Strategy Guide | 重試策略指南

StrategyWhen to UseConfig使用時機
ImmediateRare glitches, fast opsmax 2 retries罕見故障、快速操作
Exponential BackoffNetwork/API callsbase 1s, max 30s網路/API 呼叫
Circuit BreakerDownstream service issuesthreshold 50%, window 60s下游服務問題
Dead LetterUnrecoverable after retriesmax 5 retries then queue重試後仍不可恢復

Checkpoint Granularity | 檢查點粒度

GranularityUse Case使用場景
Per-stepLong pipelines (> 5 steps)長管線(> 5 步驟)
Per-batchBulk data processing批量資料處理
Per-phaseMulti-phase workflows多階段工作流程
Start/EndShort, atomic operations短暫的原子操作

Workflow | 工作流程

  1. DETECT - Identify failure type and scope
  2. DIAGNOSE - Determine root cause category
  3. RECOVER - Apply appropriate recovery strategy
  4. VERIFY - Confirm system state is consistent
  5. RESUME - Continue from last checkpoint

  1. 偵測 - 識別故障類型與範圍
  2. 診斷 - 判定根本原因類別
  3. 恢復 - 套用適當的恢復策略
  4. 驗證 - 確認系統狀態一致
  5. 繼續 - 從上一個檢查點繼續

Usage Examples | 使用範例

User: /durable deploy-pipeline
AI: Analyzing deploy-pipeline for durability...
    Steps: build → test → deploy → verify
    Risk: deploy step has no rollback strategy.
    Recommendation: Add checkpoint after test,
    configure rollback for deploy using blue-green strategy.

Next Steps Guidance | 下一步引導

After /durable completes, the AI assistant should suggest:

持久性分析完成。建議下一步 / Durability analysis complete. Suggested next steps:

  • 執行 /methodology 選擇適合的開發方法論 ⭐ Recommended / 推薦 — Choose suitable methodology
  • 執行 /commit 提交持久化設計變更 — Commit durability design changes

Version History | 版本歷史

VersionDateChanges變更
1.0.02026-03-24Initial release初始版本

AI Agent Behavior | AI 代理行為

完整的 AI 行為定義請參閱對應的命令文件:/durable

For complete AI agent behavior definition, see the corresponding command file: /durable

License | 授權

CC BY 4.0

Signals

GitHub stars
73
Forks
13
Last commit
Sep 2026
Advanced
Catalog kind
skill
Gateway key
durable
Source
github.com/asiaostrich/universal-dev-standards