bmad-distillator

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

Updated Apr 9, 2026
One-click install
npx skills add https://github.com/5onyy/essential-ai-agent-skills --skill bmad-distillator-5onyy
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: bmad-distillator
Source: https://github.com/5onyy/essential-ai-agent-skills/tree/main/.cursor/skills/bmad-distillator
Command: npx skills add https://github.com/5onyy/essential-ai-agent-skills --skill bmad-distillator-5onyy

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve? Feeding large sets of source documents (briefs, discovery notes, research reports) into downstream LLM workflows wastes tokens on prose, transitions, and repetition, while naive summarization loses critical facts, decisions, and constraints. This Skill produces lossless, hyper-compressed distillates that preserve every fact while stripping human-oriented overhead. ## Core Features & Use Cases - Lossless Compression: Extracts and deduplicates every fact, decision, constraint, and relationship from source documents into dense thematic bullets, with completeness verification against source headings and named entities. - Semantic Splitting: Automatically splits large distillates into self-contained section files with an index, so downstream workflows can load only the topics they need. - Round-Trip Validation: Optionally reconstructs source documents from the distillate alone and diffs them against originals to prove no information was lost. - Use Case: Before writing a PRD, distill a product brief plus its discovery notes into a single ~1,500-token distillate that the PRD workflow consumes as sole context without information loss. ## Quick Start Distill the documents in the docs/product folder into a distillate for PRD creation with round-trip validation enabled.

Frequently Asked Questions about bmad-distillator

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

FAQPage Schema
How do I compress documents for LLM context without losing information?▼

Run the distillation workflow on your source files, which extracts every fact, decision, and constraint into dense thematic bullets. A completeness check verifies all source headings and named entities appear in the output, with up to two targeted fix passes for gaps.

What is the difference between a distillate and a summary?▼

Summaries are lossy and drop details for readability. A distillate is lossless compression: it preserves every fact, decision, rejected alternative, and constraint while stripping prose transitions, hedging, and repetition that only human readers need.

What file formats can be distilled?▼

The analysis script accepts Markdown, text, YAML, and JSON files via file paths, folder paths, or glob patterns. Folders are scanned recursively while skipping directories like node_modules, .git, and .venv.

How does document distillation handle large document sets?▼

When sources exceed roughly 15,000 tokens or three files, the workflow fans out to one compressor per document group, then merges intermediate results. Distillates over about 5,000 tokens are split into semantically coherent, self-contained section files.

How can I verify a distillate did not lose information?▼

Use the --validate flag to run round-trip validation: a separate agent reconstructs the source documents from the distillate alone, then a semantic diff against the originals produces a validation report listing gaps and hallucinations.

When should I not use lossless document distillation?▼

Avoid it when a quick human-readable overview suffices, since distillates are dense bullet formats meant for machine consumption. Round-trip validation also adds significant token cost and is intended only for critical, high-stakes documents.