flows

Orchestrate multi-step agent pipelines with history tracking and error handling.

Updated Aug 23, 2026
One-click install
npx skills add https://github.com/ayo-ooo/ayo --skill flows
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: flows
Source: https://github.com/ayo-ooo/ayo/tree/main/internal/builtin/skills/flows
Command: npx skills add https://github.com/ayo-ooo/ayo --skill flows

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Flows enables you to design, run, and manage composable agent pipelines, turning ad-hoc agent actions into repeatable workflows.

Core Features & Use Cases

  • Define multi-step pipelines that pass data between agents, validate inputs, and handle errors.
  • Run flows locally or in sandboxed environments, monitor executions, and review history.
  • Use cases include automated data processing, task orchestration, and complex decision workflows.

Quick Start

Create a new flow named my-flow and run it with sample input to see the pipeline in action.

Frequently Asked Questions about flows

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I orchestrate multi-step agent pipelines with error handling?

Agent pipeline orchestration connects multi-step agent calls with history tracking and error handling. You define structured flows that pass data between agents, validate inputs, and manage execution across local or sandboxed environments.

What is the best way to automate task workflows using bash and agents?

Automating task workflows requires bash compatibility and a JSON-structured metadata format. This approach lets you define pipelines that pass data between agents, execute locally or in sandboxes, and monitor every run.

Can I run agent pipelines in a sandboxed environment and track execution history?

Yes, agent pipelines can run in sandboxed environments. The orchestration tool monitors executions and maintains history tracking, allowing you to review past runs and handle errors across both local and sandboxed agent calls.

Do I need bash compatibility and JSON metadata to automate agent workflows?

Yes, bash compatibility and a JSON-structured metadata format are required. These prerequisites integrate with the orchestration tooling to define and execute multi-step agent pipelines that pass data and handle errors.

When should I use agent orchestration pipelines instead of ad-hoc agent actions?

Use agent orchestration pipelines when you need to turn ad-hoc agent actions into repeatable workflows. This is ideal for automated data processing, task orchestration, and complex decision workflows requiring input validation and error handling.

What are the limitations of running multi-step agent flows locally?

Running multi-step agent flows locally requires bash compatibility and structured JSON metadata. While local execution supports history tracking and error handling, complex decision workflows may require sandboxed environments for safe execution.