Select pattern based on dependency shape and risk profile.
Scaffold config via scripts/workflow_scaffolder.py.md.
Define handoff contract fields for every edge.
Add retry/timeouts and output validation gates.
Dry-run with small context budgets before scaling.
Note: Bundled scripts ship as Markdown reference (.md) — copy the code out of the .md file to run it.
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.
Creator: Engineering
License: MIT
Source Repo:neekware/dojo-skillsSource Bucket:engineeringOriginal Path:engineering/agent-workflow-designer
1---2name: agent-workflow-designer-23description: Agent Workflow Designer4---5# Agent Workflow Designer67**Tier:** POWERFUL 8**Category:** Engineering 9**Domain:** Multi-Agent Systems / AI Orchestration1011---1213## Overview1415Design production-grade multi-agent workflows with clear pattern choice, handoff contracts, failure handling, and cost/context controls.1617## Core Capabilities1819- Workflow pattern selection for multi-step agent systems20- Skeleton config generation for fast workflow bootstrapping21- Context and cost discipline across long-running flows22- Error recovery and retry strategy scaffolding23- Documentation pointers for operational pattern tradeoffs2425---2627## When to Use2829- A single prompt is insufficient for task complexity30- You need specialist agents with explicit boundaries31- You want deterministic workflow structure before implementation32- You need validation loops for quality or safety gates3334---3536## Quick Start3738```bash39# Generate a sequential workflow skeleton40python3 scripts/workflow_scaffolder.py sequential --name content-pipeline4142# Generate an orchestrator workflow and save it43python3 scripts/workflow_scaffolder.py orchestrator --name incident-triage --output workflows/incident-triage.json44```4546---4748## Pattern Map4950- `sequential`: strict step-by-step dependency chain51- `parallel`: fan-out/fan-in for independent subtasks52- `router`: dispatch by intent/type with fallback53- `orchestrator`: planner coordinates specialists with dependencies54- `evaluator`: generator + quality gate loop5556Detailed templates: `references/workflow-patterns.md`5758---5960## Recommended Workflow61621. Select pattern based on dependency shape and risk profile.632. Scaffold config via `scripts/workflow_scaffolder.py.md`.643. Define handoff contract fields for every edge.654. Add retry/timeouts and output validation gates.665. Dry-run with small context budgets before scaling.6768> **Note:** Bundled scripts ship as Markdown reference (`.md`) — copy the code out of the `.md` file to run it.6970---7172## Common Pitfalls7374- Over-orchestrating tasks solvable by one well-structured prompt75- Missing timeout/retry policies for external-model calls76- Passing full upstream context instead of targeted artifacts77- Ignoring per-step cost accumulation7879## Best Practices80811. Start with the smallest pattern that can satisfy requirements.822. Keep handoff payloads explicit and bounded.833. Validate intermediate outputs before fan-in synthesis.844. Enforce budget and timeout limits in every step.8586> **Creator:** Engineering87> **License:** MIT88> **Source Repo:** `neekware/dojo-skills`89> **Source Bucket:** `engineering`90> **Original Path:** `engineering/agent-workflow-designer`
Run npx skillmds@latest add neekware/agent-workflow-designer-2 in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.
Agent Workflow Designer It is listed under AI & ML on SkillMD.
This skill has not completed SkillMD's automated safety review yet. Capability flags: docs only. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.
This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.
Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.
neekware (@neekware) published this skill. Their other Agent Skills are listed on their SkillMD profile.
neekware