Back to Skills

stream-chain

Stream-JSON chaining for multi-agent pipelines, data transformation, and sequential workflows

12,574stars1,540forksUpdated 1/21/2026

Security Assessment

Safe(100/100)
Security Score100/100

About stream-chain

The Stream-Chain skill facilitates the execution of advanced multi-step workflows by allowing the output of one agent to seamlessly flow into the next. This skill is designed to address the challenges of orchestrating complex data transformations and sequential processing pipelines, providing users with a streamlined approach to managing tasks that require collaboration among multiple agents. With Stream-Chain, users can effectively automate and manage workflows that involve intricate data manipulations, ensuring that each step builds upon the previous one for enhanced efficiency and accuracy.

Key features of the Stream-Chain skill include the ability to create custom chains using user-defined prompts, as well as the option to leverage predefined pipelines for common tasks. The custom chains offer maximum flexibility, enabling users to define each step of the workflow according to their specific needs. On the other hand, the predefined pipelines are optimized workflows that cater to frequent development tasks such as code analysis and refactoring, ensuring that users can quickly apply best practices without needing to design their own processes from scratch. These capabilities empower users to implement sophisticated multi-agent coordination through a structured data flow.

Target users for the Stream-Chain skill include software developers, data engineers, and any professionals involved in complex data processing tasks who seek to improve their workflow efficiency. Use cases range from onboarding new codebases to performing technical debt assessments and conducting code quality audits. By utilizing this skill, users can streamline their processes, reduce manual intervention, and enhance overall productivity in their development environments.

FAQ

How do I run a custom chain?

You can run a custom chain using the command 'claude-flow stream-chain run <prompt1> <prompt2> [...]'. Make sure to provide at least two prompts.

What are the requirements for using this skill?

A minimum of two prompts is required to execute a custom chain. Each prompt will be treated as a step in the workflow.

Can I use predefined pipelines for specific tasks?

Yes, predefined pipelines are available for common tasks such as analysis and refactoring. You can execute them with the command 'claude-flow stream-chain pipeline <type>'.

Is there a timeout for the steps in a custom chain?

Yes, you can set a timeout for each step using the '--timeout <seconds>' option, with a default of 30 seconds.

What options are available for debugging or verbose output?

You can enable verbose output with the '--verbose' option or enable debug mode with the '--debug' option to get detailed execution logs.

Install stream-chain

Download and extract the skill files to your .claude/skills/ directory.

Quick Setup:

  1. Copy the skill folder to .claude/skills/
  2. Claude will automatically detect and use the skill