workflow-kit

Author, validate, and execute declarative YAML workflows with shell and model nodes.

3|1|Updated Feb 21, 2026
One-click install
npx skills add https://github.com/kitaekatt/plugins-kit --skill workflow-kit-kitaekatt
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: workflow-kit
Source: https://github.com/kitaekatt/plugins-kit/tree/main/plugins/workflow-kit/skills/workflow-kit
Command: npx skills add https://github.com/kitaekatt/plugins-kit --skill workflow-kit-kitaekatt

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires openai, and includes scripts (resource) and references (resource) components.

What problem does it solve?

This skill bridges the gap between the native Workflow tool and complex automation needs by providing a structured, deterministic way to execute shell scripts and non-Claude model calls without bloating the agent's context.

Core Features & Use Cases

  • Declarative Workflows: Author complex pipelines in YAML that compile into native Workflow scripts.
  • Node Strategies: Execute deterministic shell commands or OpenRouter model calls as cheap, context-light nodes.
  • Use Case: Use this to build a multi-step pipeline that processes large datasets via shell scripts and performs intelligent classification using external models, all while maintaining a clean and efficient agent context.

Quick Start

Use the workflow-kit skill to compile and run the workflow defined in the file named project-pipeline.workflow.yaml.

Frequently Asked Questions about workflow-kit

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

FAQPage Schema
How do I execute shell scripts in declarative workflows without bloating agent context?

Declarative workflows compile YAML pipelines into native Workflow scripts, offloading deterministic shell command execution to the filesystem to maintain a clean, context-light agent state.

Can I call external models through OpenRouter inside an automation pipeline?

Yes, node strategies support executing non-Claude model calls through OpenRouter, treating them as deterministic, context-light steps within your compiled workflow pipeline.

What's the best way to build a multi-step data processing pipeline that combines shell commands and model classification?

Author a declarative YAML pipeline that leverages a standardized node contract, allowing you to chain large dataset processing via shell scripts with intelligent external model classification.

How does compiling YAML workflows into native scripts minimize token usage?

The framework uses a robust node-strategy architecture that offloads data processing to the filesystem, bypassing in-context processing to ensure efficient execution and minimal token consumption.

Do I need a specific file format to define native-integrated workflows?

Yes, you author your declarative workflow pipelines in a YAML file, such as project-pipeline.workflow.yaml, which the skill then compiles and runs natively.

Why use declarative workflows instead of directly writing shell scripts for automation?

Declarative workflows provide a structured, standardized node contract for validating and executing complex pipelines, bridging the gap between native tools and complex automation without context bloat.