agentleFS
jiahao-shao1jiahao-shao1/sjh-skills/skills/notion-lifeos/SKILL.md

notion-lifeos

Skill43 starsChanged 7 months agoMITView on GitHub

User's personal Notion LifeOS (PARA: tasks, notes, projects, Make Time journals) with pre-configured DB IDs + API access — required for any LifeOS interaction. Use for adding tasks, taking notes, journaling, querying todos, searching notes, creating projects, reflection. Triggers: 'add task', 'take a note', 'what do I need to do', 'search my notes', '加个任务', '查笔记', '待办', '记录想法', '反思', '目标偏移'. Not for Notion API docs or workspace admin.

The file

We split this skill into the rules it gives an agent

8 rules. About 0 are common, 4 are distinctive, and 4 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈4

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 4

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Read these as needed — do NOT load them all upfront:

    line 26
  2. Very specific:

    Database IDs: Read ~/.config/notion-lifeos/CONFIG.private.md for IDs (fallback: CONFIG.private.md in skill dir).

    line 67
  3. Very specific:

    Make Time dedup: Always check if today's entry exists before creating (scripts/check_today_journal.sh).

    line 69
  4. Distinctive:

    Post-op confirmation: Always confirm results to the user.

    line 72
  5. Very specific:

    Use MCP notion-fetch to read the top 15-20 most relevant notes (do NOT fetch all 50 — context budget)

    line 84
  6. Very specific:

    Use MCP notion-fetch to read those 15 notes

    line 94
  7. Distinctive:

    Select values change: NEVER hardcode — always fetch schema first

    line 125
  8. Distinctive:

    Duplicate DB names: Always use databases under the LifeOS root page

    line 126

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Notion LifeOS Skill

A Notion life management system based on the PARA method and Make Time framework.

When NOT to Use This Skill

  • Notion product questions — pricing, features, API documentation, workspace admin
  • Generic productivity advice — "how should I organize my life" without intent to store/retrieve data
  • Other note-taking apps — unless user wants to migrate data INTO Notion LifeOS
  • Database schema modification — creating new databases, adding/removing properties

Core Principles

  • Capture first, organize later — Never let processing block input
  • Proactive information surfacing — Use Relations to automatically surface information where it's needed

See JEFF_SU_SUMMARY.md for detailed design philosophy.

Reference Files

Read these as needed — do NOT load them all upfront:

FileWhen to read
references/schema.mdBefore creating/updating any entry (get field names & types)
references/mcp-guide.mdWhen using MCP tools (Claude Code / Claude.ai)
references/api-guide.mdWhen using REST API (OpenClaw / Codex / other agents)
references/query-guide.mdWhen querying/filtering data (tasks by date, done status, etc.)
references/advanced.mdFor composite intents or error handling

Operation Guide

  • With Notion MCP tools → See references/mcp-guide.md
  • With Notion API access → See references/api-guide.md
  • For structured queries (filter by date, checkbox, select) → Always use scripts or REST API. MCP notion-search is semantic only and CANNOT filter by property values.

Intent Recognition & Database Mapping

User IntentTarget DBMethod
"Take a note about XXX"NotesMCP / API
"Add a task: XXX"Taskscripts/create-task.sh / MCP / API
"The best thing today was..."Make TimeMCP / API (check dedup first)
"Create project XXX"ProjectsMCP / API
"Today's tasks"Taskscripts/query-tasks.sh --date today
"Any unfinished tasks?"Taskscripts/query-tasks.sh --undone
"Today's unfinished tasks"Taskscripts/query-tasks.sh --date today --undone
"Search notes about XX"NotesMCP notion-search
"Add a resource/reference"ResourcesMCP / API
"Query my recent thoughts"Notesscripts/query-notes.sh --type Thoughts
"List active projects"Projectsscripts/query-projects.sh
"Tasks completed this month"Taskscripts/query-tasks.sh --done --since YYYY-MM-DD --by-edited
"/challenge X"Notesscripts/query-notes.sh + CC analysis (see Reflection Commands)
"/emerge"Notesscripts/query-notes.sh + CC analysis (see Reflection Commands)
"/drift"Projects + Task + Gitscripts/collect-drift-data.sh + CC analysis (see Reflection Commands)

For Note Type selection and Make Time journal extraction logic, see references/query-guide.md. For composite intents (multiple actions in one message), see references/advanced.md.

Business Rules

  1. Database IDs: Read ~/.config/notion-lifeos/CONFIG.private.md for IDs (fallback: CONFIG.private.md in skill dir). If missing, search for "LifeOS" root page (NOT "LifeOS Template"). Last resort: guide user to references/setup.md.
  2. Schema-first: Fetch database schema before first write in a session — confirm property names and allowed select values.
  3. Make Time dedup: Always check if today's entry exists before creating (scripts/check_today_journal.sh). Update if exists.
  4. Date format: ISO-8601 only (e.g., 2026-03-08).
  5. Relations: Require target page ID — search for it first.
  6. Post-op confirmation: Always confirm results to the user.

Reflection Commands

These commands use Notion data + Git history as input, with CC performing the analysis. Output is terminal only — never write to Notion.

/challenge <topic>

Stress-test a current belief against past writing.

  1. Run scripts/query-notes.sh --type Thoughts --days 90 --limit 50
  2. Scan the returned titles for relevance to the topic
  3. Use MCP notion-fetch to read the top 15-20 most relevant notes (do NOT fetch all 50 — context budget)
  4. Analyze: identify contradictions, weak assumptions, and evidence gaps
  5. Output in Chinese: "当前信念 → 矛盾证据 → 更强的替代假设"

/emerge

Surface recurring themes and hidden patterns in recent thinking.

  1. Run scripts/query-notes.sh --type Thoughts --days 30 --limit 50
  2. Scan titles first — select top 15 most interesting/diverse
  3. Use MCP notion-fetch to read those 15 notes
  4. Identify: recurring themes, emotional patterns, unasked questions, connections the user hasn't made explicit
  5. Output in Chinese, grouped by theme, quoting original notes

/drift [N days]

Detect gaps between stated goals and actual activity.

  1. Run scripts/collect-drift-data.sh --days N (default 30)
  2. Perform semantic matching (project names like "Thinking with Image" ≠ repo names like "agentic_umm" — match by meaning, not string)
  3. Analyze:
    • Per-project: git commit count + completed task count
    • 幽灵项目: active in Notion but zero git/task activity
    • 隐性工作: heavy git activity in repos not mapped to any active project
    • Time allocation distribution
  4. Output in Chinese: "项目活跃度矩阵" + "幽灵项目" + "隐性工作" + "建议"

Gotchas (Common Pitfalls)

MCP-Specific

  • Checkbox: __YES__ / __NO__, NOT true/false
  • URL property: Prefix with userDefined: (e.g., userDefined:URL)
  • Date property: Split into date:PropName:start, date:PropName:end, date:PropName:is_datetime
  • multi_select: Comma-separated string, NOT array
  • notion-search: Cannot filter by property values — use scripts/REST API

REST API-Specific

  • database_id vs data_source_id: REST API uses database_id, MCP uses data_source_id — they differ
  • Date filter: ISO-8601 only, never natural language

General

  • Select values change: NEVER hardcode — always fetch schema first
  • Duplicate DB names: Always use databases under the LifeOS root page

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.