workflow-runner

Multi-agent YAML workflows run directly in your AI chat by parsing task dependencies and role definitions within the current LLM session, allowing the model to act as the orchestration engine without external infrastructure.

Role-based agent collaboration can be executed without external API keys by using the current session's LLM to automatically orchestrate multiple agents, such as a Product Manager or Architect, to complete complex tasks sequentially or in parallel.

Topological execution of multi-step processes is automated by providing a valid YAML workflow definition to the AI, which then steps through the roles, manages dependencies, and generates a final summary report.

Workflow orchestration works with predefined development processes by parsing agency-orchestrator YAML files, mapping out topological task dependencies, and executing the defined roles automatically within the chat session.

You do not need complex infrastructure to orchestrate multi-agent tasks; the skill uses the current LLM session as the engine, requiring only a valid YAML workflow definition and access to agent role definitions to manage state and output generation.