Kang-chen/note-organizer

NoteOrganizer Skill

This skill is responsible for automatically categorizing and storing notes in the personal knowledge base.

Workflow Steps

1. Analyze Input: Analyze the user's input to identify the main topic and relevant keywords.

2. Load Project Memory: Read project metadata files from knowledge/projects/ to understand active projects and their file organization.

   • Must read: knowledge/projects/_index.md (project index)

3. Consult Skills (Must Read):

   Important: Must use Read tool to proactively load these skill files, do not rely on auto-loading!

   Required skills (based on target directory):

   • notes-categorizer skill: Contains rules for planning, programming, biology notes
   • project-organization skill: For multi-file archival, contains project organization rules
   • knowledge-base skill: General structure and formatting
   • content-handling skill: For content integrity and logging language rules
   • For SOP/Loop pages, use the loop-page-creator skill instead

4. Identify Project: Match content keywords to active projects (see project-organization skill).

5. Determine Target Path: Based on the matching rule and project, determine the target directory:

   • With project affiliation: knowledge/{category}/{project}/
   • Without project affiliation: knowledge/{category}/

6. Generate File Name: Create a descriptive, hyphen-separated filename from the note's title (e.g., my-new-research-plan.md).

7. Format Content:

   • Add a YAML frontmatter block with creation_date, tags, and project (if applicable)
   • Ensure the content is in valid Markdown format

8. Save File: Use the edit_file tool to create a new file with the formatted content in the determined target path.

9. Update Changelog: Add entry to knowledge/logs/_changelog.md following the changelog skill format.

10. Confirm: Report the successful creation and location of the new note to the user.

Directory Structure

/knowledge/
├── projects/      # Project metadata (Agent Memory)
├── planning/      # Notes related to project planning, roadmaps, and goals.
├── programming/   # Notes related to software development, algorithms, and coding.
│   ├── {project}/       # Project-specific subdirectories
│   └── commands/        # Common commands (no project affiliation)
├── biology/       # Notes related to biology, bioinformatics, and research.
│   └── {project}/       # Project-specific subdirectories
├── inbox/         # Inbox for unprocessed notes and ideas.
├── archive/       # Immutable archive of original inbox files, organized by month.
├── _memory/       # Agent memory: structured knowledge entries extracted from archive.
│   ├── _index.md  # Knowledge index (Agent loads this first)
│   └── entries/   # Individual knowledge entry files
├── digests/       # Weekly review summaries.
├── logs/          # Logs for tracking changes and activities.
├── SOP/           # Standard Operating Procedures and company Loop documentation.
├── private/       # Private notes, credentials references, and personal information.
└── travel/        # Travel plans and related notes.

Note: For SOP/Loop pages, use the loop-page-creator skill which follows company Loop standards.

File Naming Convention

• File names should be in English, using lowercase letters
• Words should be separated by hyphens (-)
• The file extension should be .md
• Example: my-research-notes.md

Inbox Archival Workflow

1. Load project index: Read: knowledge/projects/_index.md
2. Analyze file content: Identify topics and keywords
3. Match project: Determine project affiliation based on keyword mapping
4. Load target rules: Read corresponding skill based on target directory
5. Determine target path: {category}/{project}/ or {category}/
6. Execute archival: Move file, update frontmatter, record changelog
7. Update project index: If new project, update _index.md

Weekly Review (周回顾)

Unified entry point for periodic knowledge base maintenance.

Flow

1. TODO Archive: Load notes-categorizer skill, run Archive TODO two-phase flow
2. Inbox Processing: Run three-phase inbox workflow (below)
3. Health Check: Report counts per area + overdue alerts
4. Digest: Generate knowledge/digests/_YYYY-Www-digest.md, log to changelog

Inbox Processing Workflow (Three Phases)

Phase 1 — Archive Original Files

Preserve original content integrity:

• Move each inbox/*.md file to knowledge/archive/YYYY-MM/ with date prefix
• Move associated .assets/ directories alongside their files
• Do not modify file content

Phase 2 — Extract Knowledge Entries

Create Agent-searchable memory from archived files:

• Read each archived original file
• Extract independent knowledge entries by type:
  • finding: experimental results, analysis conclusions
  • procedure: reusable step-by-step workflows
  • decision: decisions made and their rationale
  • reference: reference materials, tool configurations
• Each entry file goes to knowledge/_memory/entries/ with structured frontmatter:

---
title: "Entry title"
type: finding|procedure|decision|reference
project: ProjectName
tags: [tag1, tag2]
keywords: [keyword1, keyword2]
source: archive/YYYY-MM/original-file.md
created: YYYY-MM-DD
confidence: high|medium|low
---

• Present extraction preview to user, wait for confirmation before writing

Phase 3 — Update Indexes

• Append new entries to knowledge/_memory/_index.md table
• Update knowledge/projects/_index.md if new projects found
• Log all operations to knowledge/logs/_changelog.md

Git Auto-Commit

After each phase completes, create a git commit in the knowledge/ inner repo:

• Present commit message to user for confirmation before committing
• Format: archive: 归档 N 项TODO + 处理 M 个inbox文件 (YYYY-MM-DD)
• Only commit in knowledge/ inner repo, do not affect outer repo

Task Lifecycle

• New tasks enter 待规划任务 or 紧急任务 section
• Starting work: move to 进行中任务
• Completed: mark [x], archived during weekly review
• Cancelled: mark [~], also archived during weekly review (labeled as "已取消")
• Keep 紧急任务 under 5 items
• Evaluate 进行中任务 that exceed 2 weeks

Related Skills

• knowledge-base - Directory structure and formatting rules
• notes-categorizer - Category-specific rules (planning, programming, biology)
• project-organization - Project-based organization and rule loading
• content-handling - Content integrity and logging language rules
• changelog - Changelog format and maintenance
• private-info - Sensitive information handling
• loop-page-creator - For SOP/Loop pages