ai4pkm-helper

AI4PKM helper for onboarding guidance, quick help, and Gobi Desktop/CLI workflow integration.

16 stars

Best use case

ai4pkm-helper is best used when you need a repeatable AI agent workflow instead of a one-off prompt.

AI4PKM helper for onboarding guidance, quick help, and Gobi Desktop/CLI workflow integration.

Teams using ai4pkm-helper should expect a more consistent output, faster repeated execution, less prompt rewriting.

When to use this skill

  • You want a reusable workflow that can be run more than once with consistent structure.

When not to use this skill

  • You only need a quick one-off answer and do not need a reusable workflow.
  • You cannot install or maintain the underlying files, dependencies, or repository context.

Installation

Claude Code / Cursor / Codex

$curl -o ~/.claude/skills/ai4pkm-helper/SKILL.md --create-dirs "https://raw.githubusercontent.com/diegosouzapw/awesome-omni-skill/main/skills/cli-automation/ai4pkm-helper/SKILL.md"

Manual Installation

  1. Download SKILL.md from GitHub
  2. Place it in .claude/skills/ai4pkm-helper/SKILL.md inside your project
  3. Restart your AI agent — it will auto-discover the skill

How ai4pkm-helper Compares

Feature / Agentai4pkm-helperStandard Approach
Platform SupportNot specifiedLimited / Varies
Context Awareness High Baseline
Installation ComplexityUnknownN/A

Frequently Asked Questions

What does this skill do?

AI4PKM helper for onboarding guidance, quick help, and Gobi Desktop/CLI workflow integration.

Where can I find the source code?

You can find the source code on GitHub using the link provided at the top of the page.

SKILL.md Source

# AI4PKM Helper

This skill provides guidance for setting up and using AI4PKM (AI for Personal Knowledge Management). It helps beginners through onboarding and directs users to appropriate resources (Gobi Desktop, CLI, Orchestrator).

## AI4PKM Architecture

```mermaid
flowchart BT
    subgraph Gobi["Gobi Desktop"]
        subgraph Storage["Storage"]
            FS[File System]
        end
        subgraph Interface["Interface"]
            OBS[Obsidian]
            CC[Claude Code]
        end
        subgraph Core["Core"]
            ORC[Orchestrator]
            CLI[AI4PKM CLI]
        end
    end
    Storage --> Interface --> Core
```

**3-tier model:**
- **Storage**: Markdown files (AI-optimal format)
- **Interface**: Obsidian for viewing/editing, Claude Code for AI interaction
- **Core**: Orchestrator for automation, CLI for quick commands
- **GUI**: Gobi Desktop wraps all above with visual interface

## Learning Path

| Stage | Target | Guide |
|-------|--------|-------|
| A. PKM 초보자 | 저장 습관 만들기 | Background reading |
| B. Obsidian 입문 | 마이그레이션 + 익숙해지기 | (1) 터미널 설치 가이드 |
| C. Claude Code 입문 | CLI Agent 설치 및 첫 대화 | (2) Gobi Desktop 가이드 |
| D. 고급 사용자 | 자동화 + 커스텀 워크플로우 | (4) Orchestrator 설정 |

## When to Use This Skill

Claude should automatically load this skill when:
- User requests "AI4PKM 시작" or "온보딩 시작"
- User asks about setting up AI4PKM or Obsidian vault
- User wants to configure orchestrator or workflows
- User needs help adding tasks or understanding AI4PKM concepts
- User is new to PKM and needs guided setup

## Quick Commands

```markdown
"AI4PKM 시작" / "온보딩 시작" → Start Onboarding Module
"Orchestrator 설정" → Orchestrator Setup Module
"태스크 추가" → Task Management Module
"도움말" → Show available modules and commands
```

## Core Modules

### Module 1: Onboarding (초보자 셋업)

**Purpose**: Guide complete beginners from zero to first journal entry

**Flow** (5 steps):
1. **Assessment**: Evaluate user level (3 simple questions)
2. **Setup Vault**: Create 6 essential folders + AGENTS_beginner.md
3. **Plugin Guide**: Install 3 must-have plugins (Templater, Calendar, Dataview)
4. **First Journal**: Create today's journal interactively
5. **AI Connection**: Verify Claude Code works

**Prompts**:
- `AI4PKM - Onboarding - Step 1 - Assessment.md`
- `AI4PKM - Onboarding - Step 2 - Setup Vault.md`
- `AI4PKM - Onboarding - Step 4 - First Journal.md`
- `AI4PKM - Onboarding - Step 3 - Plugin Guide.md` (optional for MVP)
- `AI4PKM - Onboarding - Step 5 - AI Connection.md` (optional for MVP)

**Templates Used**:
- `AGENTS_beginner.md`: Simplified agent rules for beginners
- `daily_journal_template.md`: Daily journal template

### Module 2: Orchestrator Setup (워크플로우 자동화)

**Purpose**: Configure orchestrator.yaml for event-driven automation

→ **상세 가이드**: (4) Orchestrator 설정

**Key Concepts**:
- **Node**: Event-driven agent that triggers on file changes or schedules
- **Executor**: AI engine to run (claude_code, codex_cli, gemini_cli)
- **Poller**: External data sync (limitless, apple_photos, apple_notes)

**Setup Process**:
1. Explain orchestrator purpose and benefits
2. Show orchestrator.yaml structure (nodes, pollers, defaults)
3. Configure first agent (e.g., EIC - Enrich Ingested Content)
4. Test agent execution with file trigger

**Prompt**: `AI4PKM - Orchestrator Setup.md`

**Example Configuration** (current format):
```yaml
version: "1.0"

orchestrator:
  prompts_dir: _Settings_/Prompts
  tasks_dir: _Settings_/Tasks
  logs_dir: _Settings_/Logs
  skills_dir: _Settings_/Skills
  max_concurrent: 3
  poll_interval: 1

defaults:
  executor: claude_code
  timeout_minutes: 30
  max_parallel: 3
  task_create: true

nodes:
  - type: agent
    name: Enrich Ingested Content (EIC)
    input_path: Ingest/Clippings
    output_path: AI/Articles
    executor: claude_code

  - type: agent
    name: Daily Driver Workflow (DDW)
    cron: "15,45 * * * *"
    output_path: Journal
    completion_status: IGNORE
    agent_params:
      calendars:
        - Default
        - Work
      past_hours: 24
      future_days: 3
      goals_file: _Settings_/Goals & Principles.md
      roundup_lookback_days: 7

pollers:
  limitless:
    enabled: true
    target_dir: "Ingest/Limitless"
    poll_interval: 300
```

**Node Types**:

| Type | Trigger | Example |
|------|---------|---------|
| File-triggered | New/updated file in input_path | EIC, GDR |
| Cron-scheduled | Time-based (cron expression) | DDW |

**설정 필드 참조**:

| 필드 | 설명 |
|------|------|
| `name` | Agent 이름 |
| `prompt` | 프롬프트 파일 참조 (ABBR) |
| `input_path` | 입력 디렉토리 |
| `output_path` | 출력 디렉토리 |
| `executor` | 실행 엔진 (claude_code, codex_cli, gemini_cli) |
| `cron` | 스케줄 실행 (cron 표현식) |
| `workers` | Multi-Worker 설정 (여러 모델 동시 실행) |

**Multi-Worker 설정** (여러 AI 모델 비교/평가):

```yaml
- type: agent
  name: Article Summary Comparison
  prompt: EIC
  input_path: Ingest/Articles
  workers:
    - executor: gemini_cli
      label: Gemini
      output_path: AI/Summary/Gemini
    - executor: claude_code
      label: Claude
      output_path: AI/Summary/Claude
```

### Module 3: Task Management (태스크 관리)

**Purpose**: Add, update, and track tasks in AI4PKM system

**Operations**:
- **Add Task**: Create new task with description, category, priority
- **Update Status**: Mark tasks as in_progress, completed, or blocked
- **View Tasks**: List tasks by status, category, or date

**Prompt**: `AI4PKM - Add Task.md`

**Task Categories**:
- 🔍 Research
- 📐 Design
- 🛠️ Implement
- ✍️ Write
- 📝 Docs & System

## Voice Mode Configuration

### Language Support
- Support both English and Korean
- Use `tts_model="gpt-4o-mini-tts"` for natural pronunciation
- Continue conversation naturally

### Listening Behavior
- **Patient listening**: Wait for user to finish (use `min_listen_duration=8`)
- **Adaptive waiting**: Don't interrupt incomplete thoughts
- Use `vad_aggressiveness=0` for permissive voice detection

### Progress Updates
- For long operations (10+ seconds), provide voice updates
  - "지금 폴더 생성 중이에요..." / "플러그인 확인하는 중입니다..."
- Keep user informed throughout process

## User Level Detection

### Assessment Questions (Step 1)
1. **Obsidian 경험**: "옵시디언 써보신 적 있으세요?"
   - 처음 / 들어봄 / 써봄 → beginner
   - 몇 달 사용 → intermediate
   - 1년+ 사용 → advanced

2. **PKM 목적**: "노트 관리를 왜 하고 싶으세요?"
   - 메모 정리 → beginner
   - 지식 체계화 → intermediate
   - 제2의 뇌 구축 → advanced

3. **기존 시스템**: "지금 어떻게 노트 관리하세요?"
   - 없음/종이 → beginner
   - Notion/Evernote → intermediate
   - 커스텀 시스템 → advanced

### Level-Based Features

| Level | Features | Approach |
|-------|----------|----------|
| Beginner | Onboarding only | Step-by-step guidance |
| Intermediate | Onboarding + Orchestrator | Brief explanations |
| Advanced | All modules | Direct to advanced features |

**Note**: MVP focuses on beginners only. Intermediate/advanced users can skip to specific modules.

## Community Feedback Integration

커뮤니티 피드백 기반 온보딩 원칙:

→ **피드백 소스**: [[AI/Events/2025-10-30 AI4PKM 커뮤니티 미팅 by Claude Code|커뮤니티 미팅]], [[Projects/AI4PKM/AI4PKM User Interview|User Interview]]

### 핵심 이슈와 대응 전략

| 이슈 | 문제 | 대응 |
|------|------|------|
| **개념적 혼란** | AI4PKM이 뭔지 모르겠음 | 첫 안내에서 "AI로 지식 관리" 간단 설명 |
| **기술적 장벽** | 터미널/CLI가 낯설음 | No-Code 경로 강조, Step 4 선택사항 |
| **백지 공포증** | 뭐부터 해야 할지 막막함 | Step 3에서 **저널 먼저** → 빠른 성취감 |
| **습관 형성** | 2-3주 후 이탈 | 점진적 복잡도 소개 |
| **비개발자 불안** | "따라갈 수 있을까" | 친근한 말투, 이모지, 격려 메시지 |

### PKM 근본적 질문 대응

**커뮤니티에서 제기된 질문들**:

1. **"정리만 하고 학습이 안 됨"**
   - 대응: Express 단계 강조 (Week 2+에서 창작/공유 소개)
   - CODE 사이클: Capture → Organize → Distill → **Express**

2. **"측정이 어려움"**
   - 대응: 작은 성취 체크리스트로 가시화
   - 예: "첫 저널 완성! 🎉", "7개 클리핑 달성!"

3. **"개인마다 니즈 다름"**
   - 대응: Assessment에서 맞춤형 경로 제공
   - 미래: Importer Guide로 기존 데이터 마이그레이션 지원

4. **"수집은 하는데 안 봄"**
   - 대응: Journal = Single Source of Truth로 통합
   - 매일 저널 기반으로 정보 확인

### 비개발자 친화적 접근

**No-Code 경로** (터미널 없이 사용 가능):
- 웹 클리핑 → 자동 저장
- 일기 작성
- 토픽 인덱싱 (수동)
- 주간 라운드업 (수동 트리거)

**핵심 메시지**:
> "완벽하지 않아도 됩니다. 일단 시작하세요."
> "비개발자도 80%의 기능을 터미널 없이 사용할 수 있어요."

### 시간 가이드 (레벨별)

| 레벨 | 일일 시간 | 가능한 것 |
|------|----------|-----------|
| Level 1 | 5분 | 클리핑 + 일기 한 줄 |
| Level 2 | 15분 | + 토픽 정리 + AI 요약 |
| Level 3 | 30분+ | + 워크플로우 자동화 |

## Example Interactions

### Starting Onboarding
```
User: AI4PKM 시작
Agent: AI4PKM에 오신 걸 환영합니다! 🎉
        먼저 몇 가지만 여쭤볼게요.
        옵시디언 써보신 적 있으세요?
User: 처음이에요
Agent: 좋아요! 처음부터 차근차근 안내해드릴게요.
       [Step 1 프롬프트 실행]
```

### Completing Onboarding
```
Agent: 축하합니다! 첫 저널도 만들었고, AI 연동도 확인했어요! ✅

       AI4PKM 온보딩이 완료되었습니다!
       이제 다양한 기능을 활용해보세요:
       - 클리핑 저장 및 AI 분석
       - 저널 작성 및 리뷰
       - 워크플로우 자동화 (Orchestrator)

       궁금한 점이 있으시면 언제든 물어보세요!
```

## Implementation Notes

### Prompt Chaining
Each prompt is independent but connected:
- **Input**: Previous step's output (optional)
- **Process**: Execute current step
- **Output**: Results for next step

### Semi-Automatic Execution
- Each step requires user confirmation
- Keywords: "완료", "다음", "계속" → proceed
- Keywords: "중단", "나중에" → pause
- User can resume from any step

### Error Handling
- Clear error messages in Korean
- Suggest fixes or alternatives
- Allow retry or skip (when safe)
- Never leave user stuck

## Files Structure

```
_Settings_/Skills/ai4pkm-helper/
├── SKILL.md                    # This file
└── modules/                    # (Future: detailed module guides)
    ├── onboarding.md
    ├── orchestrator.md
    └── tasks.md

_Settings_/Prompts/
├── AI4PKM - Onboarding - Step 1 - Assessment.md
├── AI4PKM - Onboarding - Step 2 - Setup Vault.md
├── AI4PKM - Onboarding - Step 4 - First Journal.md
└── ...

_Settings_/Templates/
├── AGENTS_beginner.md
├── daily_journal_template.md
└── orchestrator_template.yaml
```

## References

### AI4BM Theory Documents
- [Second Brain Concept](obsidian://open?vault=AI4BM&file=Theory/Concepts/Second%20Brain)
- [PKM Framework](obsidian://open?vault=AI4BM&file=Theory/Concepts/PKM%20Framework)
- [AI4PKM Theory Series](obsidian://open?vault=AI4BM&file=Theory/AI4PKM/Overview)

### Design Documents
- `[[AI/Analysis/2025-12-10 AI4PKM Onboarding Agent Design - Claude Code]]`
  - Original 4-week design (comprehensive reference)

## Troubleshooting

### Common Issues

**"폴더가 이미 있어요"**
→ 괜찮습니다! 기존 파일은 건드리지 않고 필요한 것만 추가할게요.

**"플러그인 설치가 안 돼요"**
→ Obsidian 설정 → 커뮤니티 플러그인 → "제한 모드 해제" 확인해주세요.

**"저널 템플릿이 안 보여요"**
→ Templater 플러그인 설정에서 템플릿 폴더를 `_Settings_/Templates`로 지정해주세요.

**"음성 모드가 안 돼요"**
→ 텍스트로도 괜찮아요! 음성은 선택 사항입니다.

## Future Enhancements

### Post-MVP (v1.1+)
- Notion migration assistant
- Best practices guide module
- Advanced workflow configuration
- Habit tracking integration
- Community resource links

### Long-term Vision
- Multi-vault setup support
- Team collaboration setup
- Custom workflow builder
- Performance analytics
- Mobile setup guide

## Success Criteria

### Onboarding Success
- [ ] User completes all 5 steps within 30 minutes
- [ ] Vault structure created correctly
- [ ] First journal entry exists
- [ ] User can interact with Claude Code naturally

## Notes

- **Focus on simplicity**: MVP for beginners only
- **Voice-friendly**: Design for conversational interaction
- **Quick wins**: Each step provides immediate value
- **Gradual learning**: Don't overwhelm with features

Related Skills

basilica-cli-helper

16
from diegosouzapw/awesome-omni-skill

This skill should be used when users need to rent GPUs, run ML training jobs, or manage compute resources on Basilica's decentralized GPU marketplace. Use it for PyTorch/TensorFlow training, distributed training setup, GPU rental management, cost monitoring, or any Basilica CLI workflows. Includes workaround for non-TTY environments like Claude Code.

Git Commit Helper

16
from diegosouzapw/awesome-omni-skill

Creates well-formatted conventional commits with intelligent change analysis. Use when creating commits, committing changes, staging files, or when the user mentions "commit", "git commit", or wants to save their work to version control. Analyzes diffs to suggest splitting commits when multiple concerns are detected.

esphome-config-helper

16
from diegosouzapw/awesome-omni-skill

This skill should be used when the user asks to "create an esphome config", "set up an esp32 device", "configure esphome yaml", "add a sensor to esphome", "fix esphome compile error", or mentions "gpio pin assignment", "wifi setup", "ota update", or error messages like "Unknown platform", "GPIO already in use", "Could not compile", or "WiFi connection failed". Provides rapid ESPHome configuration generation, troubleshooting, and validation.

bgo

10
from diegosouzapw/awesome-omni-skill

Automates the complete Blender build-go workflow, from building and packaging your extension/add-on to removing old versions, installing, enabling, and launching Blender for quick testing and iteration.

Coding & Development

large-data-with-dask

16
from diegosouzapw/awesome-omni-skill

Specific optimization strategies for Python scripts working with larger-than-memory datasets via Dask.

langsmith-fetch

16
from diegosouzapw/awesome-omni-skill

Debug LangChain and LangGraph agents by fetching execution traces from LangSmith Studio. Use when debugging agent behavior, investigating errors, analyzing tool calls, checking memory operations, or examining agent performance. Automatically fetches recent traces and analyzes execution patterns. Requires langsmith-fetch CLI installed.

langchain-tool-calling

16
from diegosouzapw/awesome-omni-skill

How chat models call tools - includes bind_tools, tool choice strategies, parallel tool calling, and tool message handling

langchain-notes

16
from diegosouzapw/awesome-omni-skill

LangChain 框架学习笔记 - 快速查找概念、代码示例和最佳实践。包含 Core components、Middleware、Advanced usage、Multi-agent patterns、RAG retrieval、Long-term memory 等主题。当用户询问 LangChain、Agent、RAG、向量存储、工具使用、记忆系统时使用此 Skill。

langchain-js

16
from diegosouzapw/awesome-omni-skill

Builds LLM-powered applications with LangChain.js for chat, agents, and RAG. Use when creating AI applications with chains, memory, tools, and retrieval-augmented generation in JavaScript.

langchain-agents

16
from diegosouzapw/awesome-omni-skill

Expert guidance for building LangChain agents with proper tool binding, memory, and configuration. Use when creating agents, configuring models, or setting up tool integrations in LangConfig.

lang-python

16
from diegosouzapw/awesome-omni-skill

Python 3.13+ development specialist covering FastAPI, Django, async patterns, data science, testing with pytest, and modern Python features. Use when developing Python APIs, web applications, data pipelines, or writing tests.

kramme:agents-md

16
from diegosouzapw/awesome-omni-skill

This skill should be used when the user asks to "update AGENTS.md", "add to AGENTS.md", "maintain agent docs", or needs to add guidelines to agent instructions. Guides discovery of local skills and enforces structured, keyword-based documentation style.