stream-chain

Multi-agent workflow orchestration chains AI agent outputs sequentially, with each step receiving the complete result from the previous one. stream-chain automates this by streaming outputs across custom chains and predefined pipelines, enabling complex tasks like code analysis, security auditing, and data transformation in a single automated sequence.

Yes. stream-chain supports predefined pipelines for security audits, code refactoring, testing, and optimization. It streams each agent's output to the next, maintaining context throughout. For example, run a security workflow that analyzes vulnerabilities, categorizes issues, proposes fixes, and generates test cases automatically.

Custom chains let you define flexible sequences of prompts tailored to your specific workflow. Predefined pipelines are battle-tested workflows for common tasks like analysis, refactoring, and optimization. Both support context flow and streaming, so choose custom chains for unique requirements and pipelines for standard tasks.

stream-chain automatically propagates per-step context, passing the complete output from each agent as input to the next. This ensures continuity across your workflow. Configure verbose, timeout, and debug options to control execution behavior and monitor context flow through your pipeline.

Yes. stream-chain handles complex data transformations and sequential multi-agent processing by streaming outputs between steps. Use run mode for custom chains or pipeline mode for predefined workflows. Both support configurable options and maintain full context across all steps.

stream-chain depends on claude-flow as its foundation. It builds on claude-flow to add multi-agent orchestration and streaming output chaining, enabling sophisticated workflows where agents collaborate sequentially with continuous context propagation.