bmad-distillator

Compress source documents into lossless distillates for LLM consumption.

Updated Sep 27, 2025

Search Tags:#workflow #llm #compression #document-processing #token-efficiency #information-extraction #distillation

Ready to run

npx skills add https://github.com/Cbanzaime23/Booking-System --skill bmad-distillator-cbanzaime23

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: bmad-distillator
Source: https://github.com/Cbanzaime23/Booking-System/tree/main/.agent/skills/bmad-distillator
Command: npx skills add https://github.com/Cbanzaime23/Booking-System --skill bmad-distillator-cbanzaime23

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

Distiller converts multiple source documents into a single lossless distillate that preserves all facts, decisions, constraints, and relationships while minimizing context overhead for downstream LLM workloads.

Core Features & Use Cases

Lossless compression of documents to maximize token efficiency without data loss
Supports single distillates or semantically split multi-part outputs with section manifests
Handles facts, decisions, constraints, open questions, and named entities across document types
Suitable for product briefs, discovery notes, architecture docs, and research reports
Integrates with a compression pipeline including analysis, compression, verification, and optional round-trip validation

Quick Start

Provide the source documents and optional downstream_consumer to generate a lossless distillate.

Frequently Asked Questions about bmad-distillator

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I compress multiple documents for LLM consumption without losing facts?▼

To compress documents for LLM consumption without losing facts, apply a lossless distillation process that analyzes, deduplicates, and extracts facts, decisions, and constraints into a single optimized output. This preserves all critical relationships while minimizing context overhead.

How do I semantically split a large distillate output for better LLM context handling?▼

To semantically split a distillate output, configure the compression process to generate multi-part outputs with section manifests. This breaks the single compressed document into logical semantic sections, optimizing context delivery for downstream LLM workloads.

Can I process mixed document types like product briefs and architecture docs together?▼

Yes, you can process mixed document types like product briefs, discovery notes, architecture docs, and research reports together. The distillation pipeline handles facts, decisions, constraints, open questions, and named entities across diverse document formats.

How does round-trip validation work during document distillation?▼

Round-trip validation during document distillation verifies that the compressed distillate accurately preserves information from the source documents. This optional step follows the analysis, compression, and verification pipeline to ensure lossless information extraction.

What is the best way to reduce token overhead when feeding multiple reports to an LLM?▼

The best way to reduce token overhead when feeding multiple reports to an LLM is lossless document distillation. It converts multiple source documents into a single compressed distillate, applying deduplication and rule-based compression to maximize token efficiency without data loss.