bmad-distillator

Compress source documents into lossless token-efficient distillates for LLM pipelines.

Updated Feb 11, 2026

Search Tags:#documentation #llm #compression #knowledge-management #token-efficiency #distillation #lossless

Ready to run

npx skills add https://github.com/mschuerig/peach-ios --skill bmad-distillator-mschuerig

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: bmad-distillator
Source: https://github.com/mschuerig/peach-ios/tree/main/_bmad/core/skills/bmad-distillator
Command: npx skills add https://github.com/mschuerig/peach-ios --skill bmad-distillator-mschuerig

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

Distillator produces hyper-compressed, lossless distillates from a set of source documents. A distillate preserves every fact, decision, constraint, and relationship while stripping overhead humans need to load into downstream workflows.

Core Features & Use Cases

Produces a single dense distillate or semantically segmented section distillates for downstream LLM workflows.
Preserves core facts, decisions, constraints, and relationships; removes prose, hedging, and formatting to minimize token usage.
Use cases include PRD creation, architecture planning, knowledge consolidation, and long-term archival-ready context for LLMs.

Quick Start

Provide source_documents and optional downstream_consumer to generate a 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 source documents for LLM context without losing facts?▼

You compress source documents into a lossless distillate by stripping prose, hedging, and formatting overhead while preserving every fact, decision, constraint, and relationship. This minimizes token usage for downstream LLM workflows.

What is lossless distillation for knowledge management?▼

Lossless distillation is the process of compressing documents into a dense format that retains all core facts and relationships. It removes human-centric overhead to produce hyper-compressed, LLM-ready context for knowledge consolidation.

How do I generate a distillate for PRD creation or architecture planning?▼

You generate a distillate by providing source documents and an optional downstream consumer. The process applies stage-driven processing and semantic splitting to output either a single dense distillate or semantically segmented sections.

Does the distillation process support semantically segmented section distillates?▼

Yes, the distillation process supports semantic splitting to produce semantically segmented section distillates. This allows you to organize hyper-compressed context into specific segments that match downstream LLM pipeline requirements.

What is the best way to reduce token usage when feeding large document sets to LLMs?▼

The best way to reduce token usage is to compress large document sets into a lossless distillate. This strips formatting and prose while strictly preserving facts and relationships, ensuring token efficiency for downstream LLM workflows.

Can I validate the distillate format before using it in downstream pipelines?▼

Yes, the distillation process includes optional validation alongside strict distillate-format rules. This ensures the generated output satisfies frontmatter metadata requirements and preserves structural integrity for downstream LLM pipelines.