bmad-advanced-elicitation

Iterative content refinement improves LLM output by applying a deterministic elicitation flow to recent drafts. It loads methods from a registry to systematically reconsider and enhance content, returning an auditable, improved version.

Structured elicitation for content review is a workflow that enforces an exact step order to refine documents. It uses an agent-manifest and methods registry to drive iterative enhancement, ensuring structured escalation and provenance tracking.

To apply an elicitation workflow to an existing document, the process loads a methods.csv and agent-manifest to drive the flow. It enforces a strict step order, applying selected methods to the current draft to produce an enhanced content version.

Yes, a methods registry and agent-manifest are required to enforce the deterministic elicitation flow. These components drive the step order and apply the selected refinement methods to preserve provenance and create auditable versions.

Iterative elicitation applies to content review and refinement tasks across both prompts and documents. It performs structured escalation wherever a deterministic flow is needed to improve recent LLM output and generate enhanced versions.

The best way to ensure provenance during LLM content refinement is using a deterministic elicitation workflow. It preserves provenance by applying registered methods in an exact step order, creating auditable versions of the enhanced content.