Agent 工作流设计

设计与多媒体社区

作者仓库星标 15,974
作者仓库 claude-skills

解读按原文结构重写，命令、链接、术语均保留；右侧可核对作者原始 SKILL.md

方法与流程

Quick Start：原文以此为独立章节。
Recommended Workflow：Select pattern based on dependency shape and risk profile. Scaffold config via scripts/workflowscaffolder.py. Define handoff contract fields for every edge.

适用与边界

When to Use：A single prompt is insufficient for task complexity You need specialist agents with explicit boundaries You want deterministic workflow structure before implementation
Common Pitfalls：Over-orchestrating tasks solvable by one well-structured prompt Missing timeout/retry policies for external-model calls Passing full upstream context instead of targeted artifacts

原文中的明确线索

要点：「Tier」、「Category」、「Domain」、「Design production-grade multi-agent workflows with clear pattern choice, handoff contracts, failure handling, and cost/context controls.」、「1. Select pattern based on dependency shape and risk profile.」、「1. Start with the smallest pattern that can satisfy requirements.」
文件与命令：sequential、parallel、router、orchestrator、evaluator、references/workflow-patterns.md、scripts/workflowscaffolder.py

流狐整理：以上内容来自当前 SKILL.md 的章节与原词；未补写作者没有声明的工具、兼容性或能力。

Agent Workflow Designer

Tier: POWERFUL
Category: Engineering
Domain: Multi-Agent Systems / AI Orchestration

Overview

Design production-grade multi-agent workflows with clear pattern choice, handoff contracts, failure handling, and cost/context controls.

Core Capabilities

Workflow pattern selection for multi-step agent systems
Skeleton config generation for fast workflow bootstrapping
Context and cost discipline across long-running flows
Error recovery and retry strategy scaffolding
Documentation pointers for operational pattern tradeoffs

When to Use

A single prompt is insufficient for task complexity
You need specialist agents with explicit boundaries
You want deterministic workflow structure before implementation
You need validation loops for quality or safety gates

Quick Start

# Generate a sequential workflow skeleton
python3 scripts/workflow_scaffolder.py sequential --name content-pipeline

# Generate an orchestrator workflow and save it
python3 scripts/workflow_scaffolder.py orchestrator --name incident-triage --output workflows/incident-triage.json

Pattern Map

sequential: strict step-by-step dependency chain
parallel: fan-out/fan-in for independent subtasks
router: dispatch by intent/type with fallback
orchestrator: planner coordinates specialists with dependencies
evaluator: generator + quality gate loop

Detailed templates: references/workflow-patterns.md

Recommended Workflow

Select pattern based on dependency shape and risk profile.
Scaffold config via scripts/workflow_scaffolder.py.
Define handoff contract fields for every edge.
Add retry/timeouts and output validation gates.
Dry-run with small context budgets before scaling.

Common Pitfalls

Over-orchestrating tasks solvable by one well-structured prompt
Missing timeout/retry policies for external-model calls
Passing full upstream context instead of targeted artifacts
Ignoring per-step cost accumulation

Best Practices

Start with the smallest pattern that can satisfy requirements.
Keep handoff payloads explicit and bounded.
Validate intermediate outputs before fan-in synthesis.
Enforce budget and timeout limits in every step.

流狐档案作者与许可取自来源；运行、权限和网络为流狐检测或估算

流狐分类: 设计与多媒体
作者声明 Agent: 未找到明确声明；不据此推断已兼容或已测试
静态检查: 88 / 100 · 启发式扫描，不代表运行安全
作者 / 版本 / 许可: @alirezarezvani · 未声明 license
流狐 Token 估算: 低消耗
流狐接入估算: 即装即用
是否需要外部 API Key: 未发现要求
检测到的系统要求: 未声明
底层运行要求: 未声明
检测到的文件与系统行为: 只读
允许写入 / 修改
检测到的网络行为: 仅限本地
安装命令数: 无（仅作为资料）

档案由构建时根据 SKILL.md 与安装命令自动衍生，可能与作者实际意图存在差异。

需要注意： 未限定 allowed-tools，默认拥有全部工具权限。

输出预览 agent-workflow-designer.preview

作者没有在当前 SKILL.md 中定义固定输出样例。

SKILL.md SKILL.md

# Agent Workflow Designer

**Tier:** POWERFUL  
**Category:** Engineering  
**Domain:** Multi-Agent Systems / AI Orchestration

---

## Overview

Design production-grade multi-agent workflows with clear pattern choice, handoff contracts, failure handling, and cost/context controls.

## Core Capabilities

- Workflow pattern selection for multi-step agent systems
- Skeleton config generation for fast workflow bootstrapping
- Context and cost discipline across long-running flows
- Error recovery and retry strategy scaffolding
- Documentation pointers for operational pattern tradeoffs

---

## When to Use

- A single prompt is insufficient for task complexity
- You need specialist agents with explicit boundaries
- You want deterministic workflow structure before implementation
- You need validation loops for quality or safety gates

---

## Quick Start

```bash
# Generate a sequential workflow skeleton
python3 scripts/workflow_scaffolder.py sequential --name content-pipeline

# Generate an orchestrator workflow and save it
python3 scripts/workflow_scaffolder.py orchestrator --name incident-triage --output workflows/incident-triage.json
```

---

## Pattern Map

- `sequential`: strict step-by-step dependency chain
- `parallel`: fan-out/fan-in for independent subtasks
- `router`: dispatch by intent/type with fallback
- `orchestrator`: planner coordinates specialists with dependencies
- `evaluator`: generator + quality gate loop

Detailed templates: `references/workflow-patterns.md`

---

## Recommended Workflow

1. Select pattern based on dependency shape and risk profile.
2. Scaffold config via `scripts/workflow_scaffolder.py`.
3. Define handoff contract fields for every edge.
4. Add retry/timeouts and output validation gates.

…

证据边界与执行链路

作者原文负责流程事实；流狐只索引当前章节、要点、文件与命令。

章节    -> Overview → Core Capabilities → When to Use → Quick Start → Pattern Map → Recommended Workflow
要点    -> Tier · Category · Domain · Design production-grade multi-agent workflows with clear pattern choice, handoff contracts, failure handling, and cost/context controls. · 1. Select pattern based on dependency shape and risk profile. · 1. Start with the smallest pattern that can satisfy requirements.
文件/命令 -> sequential · parallel · router · orchestrator · evaluator · references/workflow-patterns.md · scripts/workflowscaffolder.py
内容 SHA-256 -> e85adb0994b8