AgentSkillsCN

note-refining-skills

一项用于精炼、重组并改进笔记的技能,让笔记更加清晰、更具可操作性。

SKILL.md
--- frontmatter
name: note-refining-skills
description: A skill for refining, restructuring, and improving notes to make them clearer and more actionable.

Note Refining Skill

Overview

This skill helps AI agents refine raw notes, meeting minutes, brainstorming dumps, and other unstructured text into well-organized, clear, and actionable documents.

When to Use

  • When the user asks to clean up, refine, or restructure their notes
  • When processing raw meeting minutes or brainstorming outputs
  • When converting unstructured text into structured formats

Instructions

Step 1: Analyze the Input

Read the raw note content and identify:

  • Key topics — the main subjects discussed
  • Action items — tasks or follow-ups mentioned
  • Decisions — conclusions or agreements reached
  • Open questions — unresolved items

Step 2: Restructure

Reorganize the content into a clear structure:

  1. Summary — A 1-2 sentence overview
  2. Key Points — Bullet points of main topics, grouped by theme
  3. Action Items — Each with owner (if mentioned) and deadline (if mentioned)
  4. Decisions — What was decided and the rationale
  5. Open Questions — Items that need follow-up

Step 3: Refine Language

  • Fix grammar and spelling
  • Replace vague language with precise wording
  • Remove redundancy and filler
  • Maintain the original tone and intent

Step 4: Output

Produce the refined note in clean Markdown format, preserving any original context that would be lost through over-simplification.

Output Format

markdown
# [Note Title]

## Summary
[1-2 sentence overview]

## Key Points
- [Point 1]
- [Point 2]

## Action Items
- [ ] [Task] — @[Owner] (Due: [Date])

## Decisions
- [Decision 1]: [Rationale]

## Open Questions
- [Question 1]

Guidelines

  • Preserve intent — Never change the meaning of the original content
  • Be concise — Remove noise, keep signal
  • Stay neutral — Do not add opinions or assumptions not present in the original
  • Ask for clarification — If the original is ambiguous, ask the user before assuming