chat-summary

Convert conversations into self-contained Obsidian notes with controlled category and tag vocabulary.

Updated Apr 28, 2026
One-click install
npx skills add https://github.com/byron1st/personal-harness --skill chat-summary-byron1st
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: chat-summary
Source: https://github.com/byron1st/personal-harness/tree/main/skills/chat-summary
Command: npx skills add https://github.com/byron1st/personal-harness --skill chat-summary-byron1st

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Valuable research findings, design decisions, and technical details from AI conversations get buried and lost. This Skill turns a conversation into a single self-contained Obsidian markdown note that stays useful months later, while preventing tag and category sprawl that breaks vault search. ## Core Features & Use Cases - Structured note generation: Produces markdown with exactly five YAML frontmatter fields (title, category, tags, description, created_at) plus scoped body sections for background, findings, conclusions, open questions, and references. - Controlled taxonomy: Reuses the vault's existing category and tag vocabulary from references/taxonomy.md, with alias normalization and a reuse test before minting new tags. - Faithful summarization: Preserves specifics verbatim (versions, figures, URLs, commands), adds nothing the conversation did not contain, and separates established facts from speculation. - Use Case: After a long research session comparing Loki retention options, ask for a note and receive a ready-to-save markdown file tagged consistently with the rest of your vault, plus any new tag lines to paste back into the taxonomy. ## Quick Start Ask the assistant to summarize this conversation into an Obsidian note and save it to your vault.

Frequently Asked Questions about chat-summary

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

FAQPage Schema
How do I save an AI chat conversation as an Obsidian note?▼

Ask the assistant to summarize the conversation into a note. The Skill produces a markdown file with YAML frontmatter (title, category, tags, description, created_at) and a structured body covering background, findings, conclusions, open questions, and references.

How do I keep Obsidian tags consistent across notes?▼

The Skill reads references/taxonomy.md before choosing any category or tag, reuses existing vocabulary, normalizes aliases like k8s to kubernetes, and only mints a new tag if it will plausibly apply to at least three future notes.

What frontmatter fields does the generated note include?▼

Exactly five fields in fixed order: title, category, tags, description, and created_at. The title doubles as the filename, description is one sentence of 40-90 characters, and created_at uses YYYY-MM-DD HH:mm in local time.

Can it handle both Korean and English conversations?▼

Yes. Korean conversations use the sections 배경, 정리, 결론, 남은 것, 참고, while English conversations use Context, Notes, Conclusion, Open questions, and References, matching the conversation's language throughout the body.

When should a conversation not be turned into a note?▼

When it produced nothing worth keeping, such as a quick lookup or a one-line fix. The Skill instructs the assistant to say so plainly rather than manufacture a thin note that clutters the vault.