This paper introduces the Opus Prompt Intention Framework, designed to improve complex workflow generation using command-tuned large-scale language models (LLMs). We propose an intermediate intent capture layer between user queries and workflow generation. This layer, called the Opus Workflow Intention Framework, extracts workflow signals from user queries, interprets them into structured workflow intent objects, and generates workflows based on these intents. Our research demonstrates that this layer enables the LLM to generate logical and meaningful outputs that reliably scale with increasing query complexity. Applying the Opus Prompt Intention Framework to workflow generation consistently improves semantic workflow similarity metrics on a synthetic benchmark of 1,000 multi-intent query-workflow pairs. In this paper, we introduce the Opus Prompt Intention Framework by applying the concepts of workflow signals and workflow intents to LLM-based workflow generation. We present a reproducible and customizable LLM-based intent capture system for extracting workflow signals and workflow intents from user queries. Finally, we provide empirical evidence that the proposed system significantly improves the quality of workflow generation compared to direct generation from user queries, especially for mixed intent elicitation.
