What problem does it solve?
This skill solves the problem of information loss and token inefficiency when preparing large sets of source documents for LLM consumption. It creates dense, lossless distillates that preserve every critical fact, decision, and constraint while removing human-readable overhead.
Core Features & Use Cases
- Lossless Compression: Transforms verbose documentation into dense, structured bullet points without losing semantic depth.
- Semantic Splitting: Automatically manages token budgets by splitting large document sets into coherent, self-contained sections.
- Round-Trip Validation: Includes an optional verification stage to ensure the distillate can reconstruct the original information accurately.
- Use Case: Use this to distill a massive repository of product briefs, discovery notes, and architecture docs into a single, high-signal context file for a PRD-writing agent.
Quick Start
Run the bmad-distillator skill on the folder ./docs to create a compressed distillate for the PRD creation workflow.
