Related Pages

Related topics: Installation Guide, Skills Overview

Overview

LifeOS Plugin is a universal skills and MCP (Model Context Protocol) integration layer for LifeOS—a personal productivity operating system powered by Convex. The plugin extends AI agents (such as Claude Code and OpenCode) with 37 workflow skills, 126 MCP tools, and 28 MCP prompts, enabling them to interact with personal productivity data including tasks, projects, contacts, health metrics, finances, coaching insights, and life design systems.

Purpose and Scope

The primary purpose of LifeOS Plugin is to bridge AI agents with LifeOS data, allowing users to leverage AI capabilities for daily planning, project management, client relationship management, health tracking, and personal development through natural language commands.

The plugin provides:

• Workflow Automation: Pre-built skills that execute complete workflows (e.g., daily planning, weekly review, sprint planning)
• Direct Data Access: MCP tools enabling CRUD operations on all LifeOS entities
• Multi-Agent Compatibility: Support for Claude Code, OpenCode, and any MCP-compatible client
• Mutating and Read Operations: Both query and modification capabilities across the productivity stack

Sources: README.md:1

Architecture

System Components

graph TD
    A[AI Agent<br/>Claude Code / OpenCode] --> B[LifeOS Plugin Skills]
    A --> C[MCP Client]
    C --> D[lifeos-mcp Server<br/>@starascendin/lifeos-mcp]
    D --> E[LifeOS Convex Backend]
    E --> F[(Convex Database)]
    
    B -->|"Invokes Skills"| A
    C -->|"MCP Protocol<br/>126 Tools, 28 Prompts"| D
    
    style F fill:#e1f5fe
    style D fill:#fff3e0
    style B fill:#e8f5e9

Data Flow

sequenceDiagram
    participant User
    participant Agent as AI Agent
    participant Skills as Plugin Skills
    participant MCP as MCP Server
    participant Convex as LifeOS Backend
    
    User->>Agent: /daily-plan "Plan my day"
    Agent->>Skills: Invoke daily-plan skill
    Skills->>MCP: get_planning_context()
    MCP->>Convex: Query tasks, habits, calendar
    Convex-->>MCP: Planning data
    MCP-->>Skills: Context object
    Skills->>Skills: Build day plan
    Skills->>MCP: apply_planning_patch(mode="day")
    MCP->>Convex: Mutate due dates, priorities
    Convex-->>MCP: Confirmation
    MCP-->>Skills: Applied changes
    Skills-->>Agent: Summary report
    Agent-->>User: Today's top 3 + changes

Installation Methods

Agent-Specific Setup

Sources: README.md:16-37

Environment Configuration

Required credentials must be configured via environment variables:

export LIFEOS_CONVEX_URL=https://your-app.convex.site
export LIFEOS_USER_ID=your-user-id
export LIFEOS_API_KEY=your-api-key

Sources: README.md:5-8

MCP Server Configuration

For MCP-compatible clients, configure the server in your MCP settings:

{
  "mcpServers": {
    "lifeos": {
      "command": "npx",
      "args": [
        "@starascendin/lifeos-mcp@latest",
        "--url", "YOUR_CONVEX_URL",
        "--user-id", "YOUR_USER_ID",
        "--api-key", "YOUR_API_KEY"
      ]
    }
  }
}

Sources: README.md:31-44

Skills Catalog

The plugin provides 37 workflow skills organized into categories. Each skill is a self-contained workflow that can be invoked via slash commands (e.g., /daily-plan).

Sources: AGENTS.md:1

Daily Workflows

Reviews

Project & Client Management

People & Relationships

Inbox & Tasks

Voice & Notes

Health & Fitness

Finance

Coaching & Development

Life Design

Advanced AI Workflows

Sources: README.md:15-50, AGENTS.md:1-20

MCP Tools (126 Total)

The MCP server exposes 126 tools providing full CRUD operations across LifeOS entities:

Data Categories

Sources: AGENTS.md:21-24

Planning Workflow Tools

graph LR
    A[get_planning_context] --> B[Build Plan]
    B --> C[apply_planning_patch]
    
    C --> D["create_issue"]
    C --> E["schedule_issue"]
    C --> F["update_issue"]
    C --> G["assign_issue_to_current_cycle"]
    C --> H["set_top_priority"]
    C --> I["update_cycle_goals"]
    C --> J["save_daily_note"]
    C --> K["save_weekly_note"]
    
    style A fill:#bbdefb
    style B fill:#c8e6c9
    style C fill:#ffe0b2

The apply_planning_patch operation supports multiple mutation types depending on the planning mode:

Sources: skills/daily-plan/SKILL.md:1-35, skills/weekly-plan/SKILL.md:1-40, skills/sprint-plan/SKILL.md:1-35

MCP Prompts (28 Total)

The plugin exposes 28 MCP prompts—same workflows as the skills, but callable via MCP protocol. Any MCP-compatible client can invoke these prompts to trigger comprehensive workflow execution without directly invoking a skill file.

Sources: AGENTS.md:25-27

Core Workflow Patterns

Planning Workflow Pattern

Most planning skills follow a consistent three-step pattern:

graph TD
    A[Step 1: Gather Context<br/>get_planning_context] --> B[Step 2: Build Plan<br/>Analyze and decide]
    B --> C[Step 3: Apply Changes<br/>apply_planning_patch]
    
    A --> A1["daily=true<br/>weekly=true<br/>currentCycle=true<br/>backlog=true<br/>habits=true<br/>dailyFields=true<br/>calendar=true<br/>voiceMemos=true"]
    
    B --> B1["Pick top 3 priorities<br/>Schedule tasks by dueDate<br/>Assign backlog to cycle<br/>Update cycle goals if needed"]
    
    C --> C1["mode=day|week|cycle<br/>dryRun=false"]

Mutating vs. Read-Only Workflows

The plugin distinguishes between:

Mutating workflows execute without asking for confirmation—the user expects the plugin to modify LifeOS data automatically.

Sources: skills/daily-plan/SKILL.md:33-36, skills/weekly-plan/SKILL.md:37-40, skills/sprint-plan/SKILL.md:33-36

PPV Life Design System

The PPV (Purpose, Principles, Values) system is a core component for long-term life design. It connects desired future states to daily execution.

PPV Data Model

graph TD
    V[Vision<br/>Future State] --> I[Identity<br/>Who you are]
    V --> P[Pillars<br/>Ongoing Systems]
    V --> PR[Projects<br/>Existing LifeOS Projects]
    
    I --> B[Beliefs]
    I --> BH[Behaviors]
    
    P --> WA[Weekly Actions<br/>Small, Concrete]
    
    PR --> TASKS[Tasks/Issues]
    
    WA --> R[Reflections<br/>Weekly Energy Check]
    R --> AD[Adjustments<br/>Feedback Loop]
    
    AD --> I
    AD --> P
    AD --> PR
    AD --> WA

PPV Operations

Sources: skills/ppv/SKILL.md:1-80

Health Integration

The plugin integrates with Oura Ring data through dedicated health tools:

Health Data Points

Sources: skills/health-weekly/SKILL.md:1-50, AGENTS.md:23

Advanced AI Workflows

LLM Council

The llm-council skill uses multiple AI models to deliberate on complex decisions:

1. Round 1: Individual model responses to the question
2. Round 2: Peer review and cross-evaluation
3. Chairman Synthesis: Final authoritative answer

Output presentation includes individual responses, rankings table, and chairman's synthesis.

Sources: skills/llm-council/SKILL.md:1-60

Blind Spot Finder

The blind-spot-finder skill uses multi-model analysis to identify unknown unknowns:

1. Gather context (working memory, habits, health, finances)
2. Build a comprehensive brief
3. Run multi-model council for blind spot detection
4. Analyze patterns and local maxima
5. User selects one blind spot to work on

Sources: skills/blind-spot-finder/SKILL.md:1-60

Daily Training Report

The daily-training-report skill generates a comprehensive daily briefing combining:

• Yesterday's Results: Habit scorecard, streaks, health scores
• Today's Game Plan: Top 3 priorities, scheduled habits
• Initiative Progress: Yearly goal tracking
• Coaching Action Items: Pending homework

Data sources include habits, health metrics, daily agenda, tasks, initiatives, and coaching items.

Sources: skills/daily-training-report/SKILL.md:1-60

Customer Success Triage

The customer-success-triage skill provides structured triage for client work:

1. Fetch client workspace
2. Review threads, meetings, notes, and open tasks
3. Drill down when needed (Beeper, Fathom, Granola, notes)
4. Classify into: New Requirements, Follow-Ups, Risks/Blockers, Already Tracked
5. Create notes or tasks as appropriate

Sources: skills/customer-success-triage/SKILL.md:1-40

Updating the Plugin

# Update the plugin repository
cd /path/to/lifeos-plugin && git pull

# Update the MCP server (auto-updates with npx @latest)
# Or pin a version in .mcp.json
# "@starascendin/[email protected]"

Sources: README.md:56-62

Summary

LifeOS Plugin transforms AI agents into comprehensive personal productivity assistants by providing:

1. 37 workflow skills covering daily planning, reviews, project management, relationships, health, finance, coaching, and life design
2. 126 MCP tools for direct CRUD operations across all LifeOS entities
3. 28 MCP prompts for programmatic workflow invocation
4. Multi-agent support via Claude Code plugin system and OpenCode native skills
5. Mutating capabilities that execute changes without confirmation for automated productivity workflows

The plugin bridges the gap between AI reasoning capabilities and personal productivity data, enabling users to manage complex workflows through natural language commands while maintaining data consistency in their personal operating system.

Related Pages

Related topics: Overview

Installation Guide

This guide covers the complete installation process for integrating LifeOS Plugin with AI agents, enabling universal skills and MCP (Model Context Protocol) access to your personal productivity OS powered by Convex.

Overview

The LifeOS Plugin provides 37 workflow skills for project management, contacts, agendas, voice notes, health tracking (Oura Ring), finance, coaching, life direction, and more. The plugin works with multiple AI agent platforms including Claude Code and OpenCode.

Sources: README.md

Prerequisites

Before installing the LifeOS Plugin, you must obtain three credentials from the LifeOS Convex deployment.

Required Credentials

Sources: AGENTS.md

Environment Variable Configuration

Set the environment variables in your shell configuration:

export LIFEOS_CONVEX_URL=https://your-app.convex.site
export LIFEOS_USER_ID=your-user-id
export LIFEOS_API_KEY=your-api-key

Alternatively, you can pass these credentials directly to the MCP server via command-line arguments or configuration files.

Sources: README.md

Installation Methods

Claude Code Installation

The simplest installation method for Claude Code users.

claude plugin add github:starascendin/lifeos-plugin

This single command installs both the skills and MCP server automatically. After installation, ensure the environment variables are set for the MCP server to authenticate.

Sources: README.md

OpenCode Installation

OpenCode reads .claude/skills/ natively. Two options are available:

Option A: Copy skills directory

git clone [email protected]:starascendin/lifeos-plugin.git /tmp/life
cp -r /tmp/life/skills .claude/skills/lifeos

Option B: Symlink skills directory

ln -s /path/to/lifeos-plugin/skills .claude/skills/lifeos

Sources: README.md

Manual MCP Server Setup

For custom agent integrations, configure the MCP server manually.

{
  "mcpServers": {
    "lifeos": {
      "command": "npx",
      "args": [
        "@starascendin/lifeos-mcp@latest",
        "--url", "YOUR_CONVEX_URL",
        "--user-id", "YOUR_USER_ID",
        "--api-key", "YOUR_API_KEY"
      ]
    }
  }
}

Sources: README.md

Version Pinning

To pin a specific MCP server version, update .mcp.json:

{
  "mcpServers": {
    "lifeos": {
      "command": "npx",
      "args": [
        "@starascendin/[email protected]",
        "--url", "YOUR_CONVEX_URL",
        "--user-id", "YOUR_USER_ID",
        "--api-key", "YOUR_API_KEY"
      ]
    }
  }
}

Sources: README.md

Plugin Architecture

graph TD
    A[AI Agent] --> B[LifeOS Plugin]
    B --> C[Skills Directory]
    B --> D[MCP Server]
    D --> E[Convex Backend]
    E --> F[LifeOS Database]
    
    C --> G[37 Workflow Skills]
    G --> H[Mutating Workflows]
    G --> I[Read-only Workflows]
    
    D --> J[126 MCP Tools]
    D --> K[28 MCP Prompts]
    
    J --> L[Projects CRUD]
    J --> M[Tasks CRUD]
    J --> N[Health Data]
    J --> O[Finance Data]
    J --> P[Contacts]

Available Skills

After installation, the following skill categories become available:

Daily Workflows

Reviews

Project & Client Management

People & Relationships

Health & Fitness

Finance

Personal Development

Sources: README.md

Skill Invocation Patterns

After installation, skills are invoked using slash commands in the agent interface.

Basic Invocation

/daily-plan
/weekly-plan "2024-01-15"
/health-check
/finance-overview

Client-Focused Skills

/project-status "ACME"
/client-brief "Acme Corp"
/meeting-prep "John"
/customer-success-triage "Acme Corp"

Planning Workflows

The planning skills (daily-plan, weekly-plan, sprint-plan) use the same underlying MCP tools:

1. Call get_planning_context with relevant include flags
2. Build the plan based on current cycle, backlog, and calendar
3. Call apply_planning_patch with dryRun=false to apply mutations

Sources: skills/daily-plan/SKILL.md Sources: skills/weekly-plan/SKILL.md Sources: skills/sprint-plan/SKILL.md

Updating the Plugin

Update Plugin Repository

cd /path/to/lifeos-plugin && git pull

Update MCP Server

The MCP server auto-updates when using @latest:

npx @starascendin/lifeos-mcp@latest

Or pin to a specific version in .mcp.json to prevent unexpected updates.

Sources: README.md

MCP Tools Overview

The installation provides access to 126 MCP tools for full CRUD operations:

Sources: AGENTS.md

Troubleshooting

Environment Variables Not Recognized

Ensure environment variables are exported in the shell where the agent runs:

# Check if variables are set
echo $LIFEOS_CONVEX_URL
echo $LIFEOS_USER_ID
echo $LIFEOS_API_KEY

MCP Server Connection Issues

Verify the Convex URL format is correct:

• Must be https://*.convex.site
• User ID must match an entry in the Convex Users table
• API key must be generated from LifeOS settings

Skills Not Available

For Claude Code: Re-run the plugin installation command For OpenCode: Verify the skills directory exists at .claude/skills/lifeos/

Related Pages

Related topics: Daily Workflows, Review Workflows, Project Management, Health Integration (Oura Ring), Coaching System

Skills Overview

Introduction

Skills are the core building blocks of the LifeOS Plugin, representing 37 pre-defined workflow skills that enable AI agents to interact with LifeOS data and execute productivity operations. Each skill is a structured instruction set that guides an AI agent through a specific workflow using the LifeOS MCP (Model Context Protocol) tools.

Sources: README.md:1

Skills serve as the human-readable workflow definitions that bridge AI agent reasoning with LifeOS data mutations. They encapsulate domain knowledge about productivity workflows—including planning, reviewing, client management, health tracking, and life design—into reusable, executable units.

Architecture

graph TD
    A[AI Agent] -->|Invokes| B[Skill /slash command]
    B -->|Loads| C[SKILL.md Definition]
    C -->|Directs| D[MCP Tools]
    D -->|CRUD Operations| E[LifeOS Convex Backend]
    E -->|Returns Data| D
    D -->|Formats Response| F[User-Facing Report]
    
    G[Alternative: MCP Protocol] -->|28 Prompts| H[Direct MCP Invocation]
    H -->|Same Tools| E

The plugin provides two complementary interfaces:

Sources: AGENTS.md:1

Skill Categories

Skills are organized into functional categories that map to different aspects of personal and professional productivity.

Daily Workflows

Core daily planning and reflection skills that drive day-to-day execution.

Sources: README.md:42-45

Reviews

Periodic reflection and planning skills for maintaining strategic alignment.

Sources: README.md:46-51

Project & Client Management

Skills for managing client relationships and project execution.

Sources: README.md:52-57

People & Relationships

Skills for managing contacts, meetings, and relationship health.

Sources: README.md:58-63

Health & Fitness

Skills integrating with Oura Ring data for health monitoring.

Sources: AGENTS.md:1

Finance

Skills for financial tracking and analysis.

Coaching & Personal Development

Skills for coaching integration and personal growth tracking.

PPV Life Design

Skills for the Purpose, Priority, Vision (PPV) life design system.

Sources: skills/ppv/SKILL.md:1

Skill Structure

Each skill follows a standardized structure defined in SKILL.md files.

Anatomy of a SKILL.md File

Related Pages

Related topics: Review Workflows, Habits & Accountability, Project Management

Daily Workflows

Overview

Daily Workflows in LifeOS provide a structured, AI-powered approach to managing day-to-day productivity. These four complementary skills form a complete daily rhythm: from morning preparation through execution to end-of-day reflection.

The workflows are designed to be mutating operations that directly modify LifeOS data through the MCP (Model Context Protocol) integration with Convex. They require no user confirmation once invoked—the system applies changes automatically based on AI-generated plans.

Sources: README.md

Architecture

Integration with LifeOS MCP

Daily Workflows leverage the LifeOS MCP server (@starascendin/lifeos-mcp) which exposes 126 tools for full CRUD operations across the LifeOS data model. All workflows communicate exclusively through this MCP interface.

graph TD
    A[User/Agent] -->|Invoke Skill| B[Daily Workflow Skill]
    B -->|MCP Tool Calls| C[LifeOS MCP Server]
    C -->|HTTP/WebSocket| D[Convex Backend]
    D -->|Real-time Sync| E[LifeOS Data Store]
    
    F[Oura Ring] -->|Health Data| D
    G[Calendar] -->|Schedule Data| D
    H[Voice Memos] -->|Audio Data| D

Sources: AGENTS.md

Workflow Data Flow

graph LR
    A[daily-standup] -->|Read Context| B[get_daily_agenda]
    A -->|Read Tasks| C[get_todays_tasks]
    A -->|Read Sprint| D[get_current_cycle]
    
    E[daily-plan] -->|Read Context| F[get_planning_context]
    E -->|Write Changes| G[apply_planning_patch]
    
    H[end-of-day] -->|Write Summary| I[save_daily_note]
    H -->|Plan Tomorrow| J[schedule_issue]
    
    K[capture] -->|Quick Create| L[create_issue]
    K -->|Route| M[assign_issue_to_current_cycle]

Daily Standup

Skill File: skills/daily-standup/SKILL.md

Purpose

The daily-standup workflow provides a concise morning briefing covering today's agenda, tasks due, and sprint progress. It is a read-only operation that does not modify any data.

MCP Tools Used

Sources: skills/daily-standup/SKILL.md

Output Format

The workflow synthesizes data into a standup-style briefing:

• Today's Focus: Top 3 priorities
• Tasks Due: Tasks due today with priority levels
• Overdue: Late tasks requiring immediate triage
• Sprint Progress: Cycle completion percentage and key statistics
• Calendar: Scheduled meetings and events

Date Parameter

When $ARGUMENTS contains a date, the workflow uses that date instead of the current date for historical or future-day standups.

Daily Plan

Skill File: skills/daily-plan/SKILL.md

Purpose

The daily-plan workflow plans the day and applies mutations to LifeOS. This is a mutating workflow that directly modifies due dates, priorities, cycle assignments, and Daily Notes.

Execution Steps

graph TD
    1[Call get_planning_context] --> 2[Build Day Plan]
    2 --> 3[Pick Top 3 Priorities]
    2 --> 4[Schedule Tasks by dueDate]
    2 --> 5[Pull Backlog into Cycle]
    2 --> 6[Update Cycle Goals if Needed]
    3 --> 7[Call apply_planning_patch]
    4 --> 7
    5 --> 7
    6 --> 7
    7 --> 8[Report Changes]

Sources: skills/daily-plan/SKILL.md

Planning Context Parameters

When calling get_planning_context, the workflow includes:

Available Mutations

Post-Execution Report

After applying changes, the workflow reports:

• Today's top 3 priorities
• Tasks created, scheduled, or reassigned
• Current cycle modifications
• Daily Note content saved

End of Day

Skill File: skills/end-of-day/SKILL.md

Purpose

The end-of-day workflow provides EOD wrap-up with completion summary and tomorrow planning. This completes the daily productivity loop by reviewing accomplishments and preparing for the next day.

Core Functions

1. Completion Review: Summarize what was accomplished today
2. Tomorrow Planning: Schedule and prioritize work for the next day
3. Daily Note Updates: Save EOD reflections and plans
4. Cycle Sync: Ensure cycle progress is accurately tracked

Sources: skills/end-of-day/SKILL.md

Capture

Skill File: skills/capture/SKILL.md

Purpose

The capture workflow enables quick capture of thoughts, tasks, or notes with automatic routing. It is designed for rapid input during the day when users encounter items that need to be tracked.

Auto-Routing Logic

The capture workflow uses AI to determine:

• Whether the capture is a task, note, or reference
• Appropriate project/cycle assignment
• Priority level based on content
• Whether it belongs in the current cycle or backlog

Available Operations

Sources: skills/capture/SKILL.md

MCP Tools Reference

All Daily Workflows depend on these core MCP tools from the 126 available:

Planning Context

Mutations

Sources: AGENTS.md

Configuration

Environment Variables

export LIFEOS_CONVEX_URL=https://your-app.convex.site
export LIFEOS_USER_ID=your-user-id
export LIFEOS_API_KEY=your-api-key

MCP Server Configuration

{
  "mcpServers": {
    "lifeos": {
      "command": "npx",
      "args": [
        "@starascendin/lifeos-mcp@latest",
        "--url", "YOUR_CONVEX_URL",
        "--user-id", "YOUR_USER_ID",
        "--api-key", "YOUR_API_KEY"
      ]
    }
  }
}

Sources: README.md

Daily Rhythm Summary

graph LR
    subgraph Morning
        A[daily-standup] --> B[daily-plan]
    end
    
    subgraph Day
        C[Execute Tasks]
        D[capture] -->|Ad-hoc| C
    end
    
    subgraph Evening
        E[end-of-day]
    end
    
    B --> C
    C --> E
    E -->|Tomorrow| B

Skill Invocation

Claude Code

claude plugin add github:starascendin/lifeos-plugin

# Then invoke:
/daily-standup
/daily-plan
/end-of-day
/capture

OpenCode

Skills are read from .claude/skills/ directory. Symlink or copy:

ln -s /path/to/lifeos-plugin/skills .claude/skills/lifeos

Related Pages

Related topics: Daily Workflows, Health Integration (Oura Ring), Coaching System

Habits & Accountability

Overview

The Habits & Accountability system in LifeOS is a core productivity feature that enables users to track daily habits, maintain streaks, and stay accountable to their personal commitments. This system integrates with the broader LifeOS ecosystem, connecting habit data to health metrics, coaching insights, and daily planning workflows.

The habit system serves three primary purposes:

1. Tracking — Recording habit completions on a daily basis
2. Streaks — Motivating consistent behavior through streak counters
3. Accountability — Providing direct, actionable feedback when habits are missed or at risk

Sources: skills/habit-check/SKILL.md:1-3

Architecture

The Habits & Accountability system consists of three interconnected layers:

graph TD
    A[User Actions] --> B[Habit Check-in]
    B --> C[Streak Calculator]
    C --> D{Streak Status}
    D -->|3+ days| E[Streak Alert]
    D -->|Broken| F[Streak Broken Alert]
    D -->|Safe| G[Normal Tracking]
    
    H[get_habits] --> I[Habit Dashboard]
    J[get_habits_for_date] --> I
    K[check_in_habit] --> B
    
    I --> L[Daily Training Report]
    I --> M[Weekly Planning]
    I --> N[Blind Spot Finder]
    
    style E fill:#ff6b6b
    style F fill:#ee5a24

Core Components

Sources: skills/habit-check/SKILL.md:1-40

Habit Check Skill

The habit-check skill is the primary interface for daily habit management. It provides an interactive daily check-in experience with direct accountability messaging.

Workflow

graph LR
    A[Call get_habits_for_date] --> B[Call get_habits]
    B --> C[Build Dashboard]
    C --> D{Habits Complete?}
    D -->|Yes| E[Celebrate]
    D -->|No| F[Flag Pending]
    F --> G{Streak >= 3?}
    G -->|Yes| H[Streak At Risk Alert]
    G -->|No| I[Normal Reminder]

Data Fetching

The skill retrieves habit data through two parallel calls:

Sources: skills/habit-check/SKILL.md:14-20

Dashboard Output

The habit dashboard presents information in four sections:

1. Today's Habits

Each habit displays:

• Icon + Name
• Status: completed / pending / skipped / incomplete
• Current streak (e.g., "🔥 12 days")

2. Completion Rate

Format: X/Y habits completed today (percentage)

3. Streak Alerts

4. Never Skip a Rep

When habits are pending:

• Direct accountability message: "You haven't done X yet today. Your streak is at Y days. Don't break it."

When all habits are done:

• Celebration message: "All habits completed. No reps skipped."

Sources: skills/habit-check/SKILL.md:22-38

Interactive Features

Users can mark habits completed during the check-in by specifying habit names in $ARGUMENTS:

Input: "mark meditation done"
Action: Calls check_in_habit for today's date

Daily Training Report Integration

The daily-training-report skill aggregates habit data as part of a comprehensive personal performance overview. This demonstrates how habit tracking connects to the broader productivity system.

Habit Data Sources

The daily training report pulls habit data from three endpoints:

Sources: skills/daily-training-report/SKILL.md:8-11

Report Structure

The habit component of the daily training report follows this structure:

graph TD
    A[Daily Training Report] --> B[YESTERDAY'S RESULTS]
    A --> C[TODAY'S GAME PLAN]
    
    B --> B1[Habit Scorecard: X/Y]
    B --> B2[Streaks Maintained]
    B --> B3[Streaks Broken]
    B --> B4[Health Scores]
    B --> B5[Day Rating]
    
    C --> C1[Top 3 Priorities]
    C --> C2[Today's Scheduled Habits]
    C --> C3[Streak Counts]

Yesterday's Results Section:

• Habit scorecard: X/Y completed with list of each habit and status
• Streaks maintained or broken (broken streaks called out explicitly)
• Health scores from Oura integration
• Day rating based on habit completion + health scores

Today's Game Plan Section:

• Top 3 priorities from agenda + top priority tasks
• Habits scheduled for today with streak counts

Sources: skills/daily-training-report/SKILL.md:13-24

MCP Tools Reference

The following MCP tools are used for habit management:

get_habits_for_date Parameters

get_habits Response

The get_habits call returns:

• All defined habits
• Completion status for the queried date
• Streak counters for each habit
• Scheduled frequency (daily, specific days, etc.)

Accountability Philosophy

The Habits & Accountability system follows a "personal trainer" philosophy:

The tone is described as: "Direct, like a personal trainer. No fluff. Celebrate wins, call out misses."

Sources: skills/habit-check/SKILL.md:38-40

Related Skills

The habit system integrates with several other LifeOS skills:

Sources: skills/daily-plan/SKILL.md:9, skills/weekly-plan/SKILL.md:9, skills/blind-spot-finder/SKILL.md:10

Data Flow

sequenceDiagram
    participant User
    participant habit-check
    participant MCP Tools
    participant LifeOS
    
    User->>habit-check: Invoke skill
    habit-check->>MCP Tools: get_habits_for_date(today)
    MCP Tools->>LifeOS: Query habit table
    LifeOS->>MCP Tools: Habit status array
    MCP Tools->>habit-check: Response
    habit-check->>MCP Tools: get_habits
    MCP Tools->>LifeOS: Query all habits with streaks
    LifeOS->>MCP Tools: Full habit list
    MCP Tools->>habit-check: Response
    habit-check->>habit-check: Build dashboard
    habit-check->>User: Display results
    
    alt User marks complete
        User->>habit-check: "mark X done"
        habit-check->>MCP Tools: check_in_habit(habitId, today)
        MCP Tools->>LifeOS: Update completion record
        LifeOS->>MCP Tools: Success
        MCP Tools->>habit-check: Confirmation
        habit-check->>User: Updated dashboard
    end

Best Practices

1. Daily Check-in — Run habit-check each morning to review and update habit status
2. Streak Awareness — Pay attention to "streak at risk" alerts to maintain momentum
3. Comprehensive View — Use daily-training-report weekly to see habit patterns in context
4. Integration — Reference habit data in planning sessions for realistic scheduling

Related Pages

Related topics: Habits & Accountability, Daily Workflows, Review Workflows

Health Integration (Oura Ring)

Overview

The Health Integration system in LifeOS plugin provides comprehensive biometric data synchronization with Oura Ring devices. This integration aggregates sleep analysis, activity tracking, readiness scoring, stress measurement, and cardiovascular health metrics into a unified health dashboard accessible through MCP (Model Context Protocol) tools.

The system serves as the health data layer for 37 workflow skills, enabling AI agents to make informed recommendations based on real-time biometric data. It transforms raw Oura Ring measurements into actionable health insights for daily planning, coaching, and personal optimization workflows.

Architecture

graph TD
    subgraph "Data Sources"
        OR[Oura Ring Device]
    end
    
    subgraph "LifeOS MCP Server"
        HMC[Health MCP Tools]
        SK[Skills Layer]
    end
    
    subgraph "Health Data Categories"
        SL[Sleep]
        AC[Activity]
        RD[Readiness]
        ST[Stress]
        HR[Heart Rate]
        RS[Resilience]
        FT[Fitness]
        OX[Oxygen]
    end
    
    subgraph "Consuming Skills"
        HC[health-check]
        HW[health-weekly]
        DT[daily-training-report]
    end
    
    OR --> HMC
    HMC --> SK
    SK --> HC
    SK --> HW
    SK --> DT
    
    HMC --> SL
    HMC --> AC
    HMC --> RD
    HMC --> ST
    HMC --> HR
    HMC --> RS
    HMC --> FT
    HMC --> OX

Sources: README.md:1-40

Available Health MCP Tools

The plugin exposes 10 dedicated health MCP tools that retrieve Oura Ring data. All tools accept a days parameter to specify the lookback window.

Health Data Retrieval Tools

Sources: skills/health-check/SKILL.md:1-30

Available Health Skills

Health Check

A quick daily health overview that pulls 7 days of biometric data.

Invocation: health-check [days]

Data Sources Called:

1. get_health_sleep — 7 days
2. get_health_activity — 7 days
3. get_health_readiness — 7 days
4. get_health_heart_rate — 7 days
5. get_health_resilience — 7 days
6. get_health_vo2_max — 7 days
7. get_health_cardio_age — 7 days

Sources: skills/health-check/SKILL.md:1-35

Output Structure:

graph LR
    A[health-check] --> B[7-Day Window]
    B --> C[Aggregate Data]
    C --> D[Trend Analysis]
    D --> E[Concise Dashboard]
    E --> F[Actionable Insights]

Health Weekly

A comprehensive 7-day rolling health review that analyzes 14 days of data for deeper trend analysis.

Invocation: health-weekly

Data Sources Called:

1. get_health_sleep — 14 days
2. get_health_activity — 14 days
3. get_health_readiness — 14 days
4. get_health_stress — 14 days
5. get_health_workouts — 14 days
6. get_health_heart_rate — 14 days
7. get_health_resilience — 14 days
8. get_health_vo2_max — 14 days
9. get_health_cardio_age — 14 days
10. get_health_spo2 — 14 days

Sources: skills/health-weekly/SKILL.md:1-45

Output Structure:

graph TD
    A[health-weekly] --> B[14-Day Window]
    B --> C[Multi-Source Aggregation]
    C --> D[Category Analysis]
    D --> E[Trend Detection]
    E --> F[Weekly Report]
    F --> G[Recovery Insights]
    F --> H[Fitness Assessment]

Daily Training Report

A comprehensive daily briefing that combines health data with habit tracking and task management.

Invocation: daily-training-report

Workflow:

graph TD
    subgraph "Data Collection"
        A1[Yesterday's Habits] --> D[Synthesize Report]
        A2[Today's Habits] --> D
        A3[All Habits] --> D
        A4[Sleep 1 day] --> D
        A5[Readiness 1 day] --> D
        A6[Activity 1 day] --> D
        A7[Today's Agenda] --> D
        A8[Today's Tasks] --> D
        A9[Initiatives] --> D
        A10[Coaching Items] --> D
    end
    
    subgraph "Synthesis"
        D --> E[YESTERDAY'S RESULTS]
        D --> F[TODAY'S GAME PLAN]
        D --> G[HABIT SCORECARD]
        D --> H[STREAK STATUS]
    end

Report Sections:

Sources: skills/daily-training-report/SKILL.md:1-50

Data Flow

sequenceDiagram
    participant User
    participant Skill
    participant MCP_Tools
    participant Convex
    participant Oura

    User->>Skill: Invoke health skill
    Skill->>MCP_Tools: Call get_health_*
    MCP_Tools->>Convex: Fetch data
    Convex->>Oura: Request sync
    Oura-->>Convex: Biometric data
    Convex-->>MCP_Tools: Processed metrics
    MCP_Tools-->>Skill: Health data array
    Skill->>Skill: Aggregate & analyze
    Skill-->>User: Formatted report

Health Metrics Reference

Sleep Metrics

Activity Metrics

Readiness Metrics

Cardiovascular Metrics

Resilience Metrics

Sources: skills/health-check/SKILL.md:15-30

Configuration

Health data requires the standard LifeOS MCP server connection configured with:

{
  "mcpServers": {
    "lifeos": {
      "command": "npx",
      "args": [
        "@starascendin/lifeos-mcp@latest",
        "--url", "YOUR_CONVEX_URL",
        "--user-id", "YOUR_USER_ID",
        "--api-key", "YOUR_API_KEY"
      ]
    }
  }
}

Or environment variables:

export LIFEOS_CONVEX_URL=https://your-app.convex.site
export LIFEOS_USER_ID=your-user-id
export LIFEOS_API_KEY=your-api-key

Sources: README.md:45-60

Usage Patterns

Quick Health Check

User: health-check
→ Returns 7-day health dashboard with scores, trends, and 2-3 insights

Extended Health Review

User: health-check 14
→ Returns 14-day health overview instead of default 7 days

Weekly Deep Dive

User: health-weekly
→ Returns comprehensive 14-day analysis with recovery insights and fitness trends

Training Integration

User: daily-training-report
→ Combines health data with habits, tasks, and coaching items for complete daily briefing

Skill Comparison

Integration with Other Systems

Health data flows into multiple LifeOS subsystems:

graph LR
    subgraph "Health Data"
        H[Oura Ring]
    end
    
    subgraph "Integrated Systems"
        D[Daily Planning]
        C[Coaching]
        PP[PPV Life Design]
        T[Task Management]
    end
    
    H --> D
    H --> C
    H --> PP
    H --> T
    
    D -->|Affects| D1[Priority Setting]
    D -->|Affects| D2[Due Dates]
    C -->|Drives| C1[Action Items]
    PP -->|Influences| PP1[Energy Levels]
    T -->|Adjusts| T1[Capacity Planning]

Daily Training Report Integration

The daily-training-report skill demonstrates deep integration, combining:

• Habit compliance data with health scores
• Streak tracking synchronized with biometric trends
• Initiative progress aligned with energy levels
• Coaching action items based on health patterns

Sources: skills/daily-training-report/SKILL.md:1-50

Best Practices

1. Use health-check for daily standups — Quick 7-day snapshot provides context without overwhelming detail

1. Reserve health-weekly for planning sessions — The 14-day window reveals trends better suited for strategic decisions

1. Incorporate health data into daily-training-report — Biometric context enhances habit coaching effectiveness

1. Review resilience metrics weekly — Sleep recovery, daytime recovery, and stress contributions reveal optimization opportunities

1. Monitor VO2 max trends monthly — Cardiovascular fitness changes slowly but meaningfully over 4-week periods

Related Pages

Related topics: Client Management, Review Workflows, Daily Workflows

Project Management

Project Management in LifeOS is a comprehensive framework for managing work across projects, sprints (cycles), initiatives, and clients. It integrates with the Convex-powered personal productivity OS to provide full CRUD operations for projects, tasks/issues, cycles, phases, clients, and people/contacts through 126 MCP tools.

Overview

The Project Management system operates on multiple levels:

Core Architecture

graph TD
    A[LifeOS Plugin] --> B[Planning Context API]
    A --> C[Planning Patch API]
    B --> D[Convex Backend]
    C --> D
    
    E[get_planning_context] --> F[daily<br/>weekly<br/>currentCycle<br/>backlog<br/>habits<br/>calendar<br/>voiceMemos]
    G[apply_planning_patch] --> H[mode: day|week|cycle]
    H --> I[Mutations]
    I --> J[create_issue<br/>schedule_issue<br/>update_issue<br/>assign_to_cycle<br/>set_top_priority<br/>update_cycle_goals<br/>save_daily_note<br/>save_weekly_note]

Planning Workflow

Step 1: Gather Planning Context

The foundation of all project management operations is get_planning_context. This single API call aggregates data across multiple dimensions:

Sources: skills/daily-plan/SKILL.md:8-18

Step 2: Apply Planning Mutations

After analyzing the context, mutations are applied via apply_planning_patch with three modes:

Project Management Skills

Sprint Planning (`/sprint-plan`)

The sprint-plan skill manages the current cycle with mutating operations:

1. Fetch planning context with current cycle and backlog
2. Build cycle plan with goal updates and backlog prioritization
3. Apply mutations with mode="cycle"

Key Mutations:

// Assign work to current cycle
assign_issue_to_current_cycle(issueId, cycleId)

// Update cycle goals
update_cycle_goals(cycleId, goals)

// Schedule work with due dates
schedule_issue(issueId, dueDate)

// Set immediate priorities
set_top_priority([issueId1, issueId2, issueId3])

Sources: skills/sprint-plan/SKILL.md:1-45

Weekly Planning (`/weekly-plan`)

Weekly planning operates at a broader scope, scheduling work across multiple days:

graph LR
    A[Week Start] --> B[Update Cycle Goals]
    B --> C[Pull Backlog Items]
    C --> D[Schedule Due Dates]
    D --> E[Set Top Priorities]
    E --> F[Save Weekly Note]

Workflow:

1. Call get_planning_context with weekStartDate
2. Update active cycle goals when needed
3. Assign backlog tasks to current cycle
4. Schedule work across the week using dueDate
5. Set near-term top priorities
6. Call apply_planning_patch with mode="week" and dryRun=false

Sources: skills/weekly-plan/SKILL.md:1-40

Daily Planning (`/daily-plan`)

Daily planning focuses on immediate execution:

Mutations Available:

• create_issue - New tasks
• schedule_issue - Due date changes
• update_issue - Status, priority, estimate, title
• assign_issue_to_current_cycle - Cycle reassignment
• set_top_priority - Today's focus
• update_cycle_goals - Active cycle changes
• save_daily_note - Write to Agenda Daily Note
• add_issue_comment - Planning rationale

Sources: skills/daily-plan/SKILL.md:1-45

Daily Standup (`/daily-standup`)

Quick briefing for daily synchronization:

graph TD
    A[Daily Standup] --> B[get_daily_agenda]
    A --> C[get_todays_tasks]
    A --> D[get_overdue_tasks]
    A --> E[get_current_cycle]
    
    B --> F[Today's Focus]
    C --> G[Tasks Due]
    D --> H[Overdue Items]
    E --> I[Sprint Progress]

Output Sections:

• Today's Focus: Top 3 priorities
• Tasks Due: Tasks due today with priority
• Overdue: Late tasks needing triage
• Sprint Progress: Cycle completion percentage and stats
• Calendar: Meetings and events

Sources: skills/daily-standup/SKILL.md:1-35

Client & Project Reporting

Project Status (`/project-status`)

Provides phase breakdown, task statistics, and blocker identification:

graph TD
    A[Project Status] --> B[Phase Breakdown]
    A --> C[Task Statistics]
    A --> D[Blocker Analysis]
    
    B --> E[Active Phases]
    C --> F[Completion %]
    D --> G[Risk Items]

Sources: AGENTS.md:1-30

Client Brief (`/client-brief`)

Full client briefing combining projects and communications:

Client Health (`/client-health`)

Dashboard across all clients for relationship health tracking.

Customer Success Triage (`/customer-success-triage`)

Workflow for triaging client requests:

1. Call get_client_success_workspace with client name
2. Review workspace output:

• recentThreads - Business chats
• recentMeetings - Fathom and Granola meetings
• notes - Client notes
• openTasks - Active issues
• projects - Client projects

1. Drill down as needed:

• Chat detail: get_beeper_thread_messages
• Meeting detail: get_fathom_meeting, get_granola_meeting
• Transcript: get_fathom_transcript, get_granola_transcript
• Note history: get_client_notes

1. Classify findings:

• New Requirements: Net-new asks
• Follow-Ups: Items waiting on response
• Risks/Blockers: Scope ambiguity, overdue work, churn risk
• Already Tracked: Existing notes or tasks

Sources: skills/customer-success-triage/SKILL.md:1-40

Initiative & Cycle Review

Initiative Review (`/initiative-review`)

Yearly goal progress by category. Tracks progress against annual initiatives with configurable year (e.g., /initiative-review 2026).

Cycle Review (`/sprint-review`)

Sprint review with rollover options for managing incomplete work between cycles.

Data Model Relationships

graph TD
    I[Initiative] -->|1:N| C[Cycle]
    I -->|1:N| P[Project]
    
    C -->|1:N| T[Task/Issue]
    P -->|1:N| T
    
    C -->|has| G[Cycle Goals]
    P -->|has| Ph[Phases]
    
    T -->|assigned to| C
    T -->|scheduled on| D[Due Date]
    
    C -->|1:N| C2[Next Cycle]
    
    Cl[Client] -->|1:N| P
    Cl -->|1:N| Co[Contact]
    Cl -->|1:N| N[Notes]

MCP Tools Summary

Sources: README.md:1-50 Sources: AGENTS.md:1-60

Best Practices

1. Always fetch context first - Use get_planning_context before any planning mutation to ensure you have the latest data
2. Set dryRun=true initially - Preview changes before applying mutations
3. Use appropriate mode - Match the mutation mode (day/week/cycle) to your planning scope
4. Save notes as artifacts - Use save_daily_note and save_weekly_note to create readable plan artifacts
5. Link existing projects - When using PPV or other systems, prefer linking to existing projects over creating duplicates

Related Pages

Related topics: Project Management, People & Relationships

Client Management

Client Management in LifeOS provides a comprehensive system for managing client relationships, tracking customer-success work, and maintaining client health across your entire portfolio. It leverages the LifeOS MCP tools to aggregate data from multiple sources including Beeper threads, Fathom/Granola meetings, notes, and open tasks into unified client workspaces.

Overview

The Client Management system consists of four primary skills that work together to provide complete client visibility and actionable workflows:

Sources: README.md:28-32

Architecture

graph TD
    A[User Request] --> B{Client Management Skills}
    
    B --> C[client-brief]
    B --> D[client-health]
    B --> E[customer-success-triage]
    B --> F[inbox-triage]
    
    C --> G[get_client_success_workspace]
    D --> H[get_all_clients]
    E --> G
    E --> I[Meeting Tools]
    E --> J[Communication Tools]
    
    G --> K[Projects, Tasks, Notes, Threads]
    I --> L[get_fathom_meeting, get_granola_meeting]
    J --> M[get_beeper_thread_messages]
    
    K --> N[apply_planning_patch]
    L --> N
    M --> N
    
    N --> O[LifeOS Convex Backend]

Core Skills

Client Brief

The client-brief skill generates a comprehensive client briefing that includes all projects, communications, and current status information for a specific client. It is invoked with the client name or ID as an argument.

Sources: AGENTS.md:24

Workflow:

1. Call get_client_success_workspace with the client identifier
2. Aggregate all related data: projects, open tasks, notes, recent threads, meetings
3. Present a structured briefing with actionable insights

Client Health Dashboard

The client-health skill provides a health dashboard across all clients, allowing you to quickly identify which client relationships need attention and which are performing well.

Sources: README.md:29

Key Metrics Tracked:

• Communication frequency
• Open task counts
• Blocked or overdue items
• Recent meeting activity
• Note activity and recency

Customer Success Triage

The customer-success-triage skill is designed for reviewing customer asks, checking whether work is already tracked, capturing requirement summaries, and deciding what should become a task.

Sources: skills/customer-success-triage/SKILL.md:1-6

Input Format:

$ARGUMENTS should contain the client name or ID, plus an optional focus area

Customer Success Triage Workflow

graph TD
    A[Start Triage: Client Name/ID] --> B[Call get_client_success_workspace]
    B --> C{Workspace Data}
    
    C --> D[recentThreads]
    C --> E[recentMeetings]
    C --> F[notes]
    C --> G[openTasks]
    C --> H[projects]
    
    D --> I{Drill Down Needed?}
    E --> I
    F --> I
    G --> I
    H --> I
    
    I -->|Threads| J[get_beeper_thread_messages]
    I -->|Fathom| K[get_fathom_meeting<br/>get_fathom_transcript]
    I -->|Granola| L[get_granola_meeting<br/>get_granola_transcript]
    I -->|Notes| M[get_client_notes]
    
    J --> N{Classification}
    K --> N
    L --> N
    M --> N
    
    N --> O[New Requirements]
    N --> P[Follow-Ups]
    N --> Q[Risks / Blockers]
    N --> R[Already Tracked]
    
    O --> S[create_issue<br/>update_issue]
    P --> S
    Q --> S
    R --> S
    
    S --> T[create_client_note<br/>update_client_note]
    T --> U[Present Triage Results]

Sources: skills/customer-success-triage/SKILL.md:7-28

Classification Framework

When triaging customer-success work, findings should be classified into four categories:

Sources: skills/customer-success-triage/SKILL.md:17-20

MCP Tools Reference

Client Workspace Tools

Meeting Integration Tools

Communication Tools

Write Operations

Sources: skills/customer-success-triage/SKILL.md:21-25

Inbox Triage Integration

The inbox-triage skill works in conjunction with Client Management to process notes into actionable tasks. When client communications result in notes that need to be converted into work items, this skill provides the bridge between captured information and tracked work.

Sources: AGENTS.md:21

Typical Flow:

1. Customer-success triage identifies new requirements
2. Client notes are created or updated with requirements
3. Inbox triage converts notes into issues
4. Issues are linked to relevant projects and cycles

Best Practices

Write Operations

Based on the triage workflow design, follow these guidelines for write operations:

1. Save durable account memory with create_client_note or update_client_note for insights, decisions, and context
2. Use create_issue or update_issue only for execution work that requires tracking
3. Prefer updating existing notes/tasks over creating duplicates
4. Do not delete anything — maintain complete audit trails of client interactions

Sources: skills/customer-success-triage/SKILL.md:22-25

Drill-Down Strategy

When triaging, follow this prioritized approach:

1. Review the workspace output first — examine recentThreads, recentMeetings, notes, openTasks, and projects
2. Drill down only when needed — not every piece of information requires deep investigation
3. Match the drill-down tool to the source — use the appropriate tool for the data type being investigated

Sources: skills/customer-success-triage/SKILL.md:8-16

Usage Examples

Generate a Client Brief

/client-brief "Acme Corp"

Check Client Health Across Portfolio

/client-health

Triage Customer Success Work

/customer-success-triage "Acme Corp"

Triage with Focus Area

/customer-success-triage "Acme Corp" --focus "billing issues"

Data Flow Summary

graph LR
    A[Beeper Threads] --> D[Client Workspace]
    B[Fathom Meetings] --> D
    C[Granola Meetings] --> D
    E[Open Tasks] --> D
    F[Projects] --> D
    G[Notes] --> D
    
    D --> H{Triage Classification}
    H --> I[New Requirements]
    H --> J[Follow-Ups]
    H --> K[Risks/Blockers]
    H --> L[Already Tracked]
    
    I --> M[Issues + Notes]
    J --> M
    K --> M
    L --> N[Reference Link]
    
    M --> O[LifeOS Backend]
    N --> O

Related Skills

• Project Management — Link client work to specific projects and phases
• Sprint Plan — Assign client work to current cycles
• Weekly Review — Include client health in periodic reviews
• Follow-ups — Track outstanding client communications

Related Pages

Related topics: Client Management

People & Relationships

The People & Relationships module within LifeOS Plugin provides a comprehensive system for managing contacts, tracking interactions, preparing for meetings, and maintaining meaningful professional and personal relationships. This module serves as the social intelligence layer of the productivity OS, connecting human interaction data with actionable workflow skills.

Overview

The People & Relationships system integrates with multiple data sources to build complete dossiers on contacts, including Beeper messaging threads, Granola meeting records with AI-generated notes, voice memos, and calendar events. The system enables agents to perform relationship maintenance tasks such as identifying neglected contacts, tracking follow-up obligations, and rapidly switching context between different clients or people.

Sources: skills/contact-lookup/SKILL.md:1-5

The core philosophy treats relationships as first-class productivity assets. Rather than siloing contact information in a traditional CRM, LifeOS weaves relationship data into daily workflows, enabling proactive relationship maintenance alongside task and project management.

Core Skills

Contact Lookup

The contact-lookup skill provides a complete dossier on any person in the system.

Entry Point: get_contact_dossier with a name query

Dossier Components:

Sources: skills/contact-lookup/SKILL.md:10-25

Workflow:

graph TD
    A[User Query: Contact Name] --> B[Call get_contact_dossier]
    B --> C{Contact Found?}
    C -->|Yes| D[Extract Data Sources]
    C -->|No| E[Request Valid Name]
    D --> F[Fetch Beeper Threads]
    D --> G[Fetch Granola Meetings]
    D --> H[Fetch Voice Memos]
    F --> I[Aggregate Dossier]
    G --> I
    H --> I
    I --> J[Present Structured Output]

Output Structure:

- **Profile**: Name, relationship type, contact info, notes
- **AI Insights**: Communication style, personality, relationship tips
- **Recent Interactions**: Last voice memos, meetings, messages sorted by recency
- **Meeting History**: Granola meetings with key takeaways
- **Chat Threads**: Beeper conversation threads

Sources: skills/contact-lookup/SKILL.md:12-25

Meeting Preparation

The meeting-prep skill aggregates relevant context before scheduled meetings with contacts.

Data Pulled:

• Contact dossier from get_contact_dossier
• Recent Beeper threads with the person
• Granola meeting notes from previous encounters
• Open tasks or projects related to the person
• Any pending follow-ups or commitments

Purpose: Ensures agents enter meetings with full context, avoiding the need to re-explain background or rediscover relationship history.

Follow-ups Tracking

The follow-ups skill identifies communication obligations across the network.

Tracking Criteria:

Output: Prioritized list of contacts requiring responses, grouped by urgency and relationship type.

Sources: AGENTS.md:25-28

Relationship Pulse

The relationship-pulse skill proactively monitors relationship health by analyzing interaction frequency against relationship type expectations.

Data Sources:

• get_people — All contacts with relationship classifications
• get_beeper_threads — Message activity and recency
• get_granola_meetings — Meeting history

Sources: skills/relationship-pulse/SKILL.md:1-10

Neglect Thresholds:

Sources: skills/relationship-pulse/SKILL.md:22-28

Output Categories:

• Reach out soon: Prioritized by relationship closeness
• Consider reconnecting: Re-engagement opportunities
• Suggested touchpoints: Quick actions like replying to old threads or scheduling catch-ups

graph TD
    A[get_people] --> B[Classify Each Contact]
    A --> C[get_beeper_threads]
    A --> D[get_granola_meetings]
    C --> E[Calculate Last Interaction]
    D --> E
    E --> F{Compare to Thresholds}
    F -->|Exceeded| G[Add to Neglected List]
    F -->|Within bounds| H[Mark as Active]
    G --> I[Prioritize by Relationship]
    I --> J[Generate Action Suggestions]

Context Switch

The context-switch skill enables rapid context loading for a specific client or person, useful when switching between active workstreams.

Usage: When working on multiple clients or projects, agents can invoke context-switch to immediately load relevant data without manual retrieval.

Loaded Data:

• Recent client notes
• Active projects and tasks
• Meeting history
• Outstanding communications
• Project-specific context

Sources: AGENTS.md:30-31

Data Models

Person Entity

Interaction Record

Relationship Health Score

MCP Tool Integration

The People & Relationships module leverages the LifeOS MCP server for data operations:

Read Operations

Sources: skills/contact-lookup/SKILL.md:16-22

Write Operations

Sources: skills/contact-lookup/SKILL.md:25-27

Workflow Integration

The People & Relationships module connects with other LifeOS workflow skills:

Integration with Planning

During daily and weekly planning, the system can surface:

• Pending follow-ups requiring attention
• Meeting preparations needed
• Relationship pulse alerts

graph LR
    A[Daily Planning] --> B[get_planning_context]
    B --> C{Include People Data?}
    C -->|Yes| D[Fetch Follow-ups]
    C -->|Yes| E[Fetch Relationship Pulse]
    D --> F[Surface in Day Plan]
    E --> F

Integration with Reviews

During cycle and initiative reviews, contact data informs:

• Client health across projects
• Communication patterns with stakeholders
• Meeting frequency analysis

Integration with Capture

The inbox-triage skill uses person data for:

• Linking captured notes to contacts via link_memo_to_person
• Suggesting contacts when notes mention people
• Creating tasks tied to relationship obligations

Sources: skills/inbox-triage/SKILL.md:10-18

Relationship Intelligence

AI-Generated Insights

The system uses AI to generate relationship intelligence:

• Communication style: How the person prefers to communicate
• Personality indicators: Extracted from interaction patterns
• Relationship tips: Customized advice for maintaining the relationship

These insights are generated during get_contact_dossier calls and stored as part of the person record.

Blind Spot Detection

The blind-spot-finder skill incorporates relationship data to identify patterns:

• Screening time analysis vs. relationship investment
• Say-do gaps in stated priorities vs. actual relationship maintenance
• Patterns the user cannot see from inside their own behavior

Sources: skills/blind-spot-finder/SKILL.md:5-12

Best Practices

Regular Maintenance

1. Run relationship-pulse weekly to identify neglected contacts
2. Review follow-ups daily during planning
3. Use contact-lookup before any significant meeting

Data Quality

1. Ensure relationship types are accurately set for proper threshold calculation
2. Add notes during interactions for future context
3. Link voice memos and meeting notes to relevant contacts

Avoiding Duplication

• Prefer updating existing notes over creating duplicates
• Use update_client_note instead of create_client_note when history exists
• Link to existing contacts rather than creating new person records

Sources: skills/contact-lookup/SKILL.md:26-27

Summary

The People & Relationships module provides a holistic approach to relationship management within LifeOS. By integrating messaging, meetings, voice notes, and AI-generated insights, it enables proactive relationship maintenance alongside productivity work. The system treats relationships as living data that requires regular attention, not static contacts to be occasionally referenced.

Key capabilities include:

• Complete contact dossiers aggregating all interaction data
• Proactive neglect detection with customizable thresholds
• Meeting preparation with full historical context
• Follow-up tracking across multiple communication channels
• Rapid context switching for multi-client workflows

Related Pages

Related topics: Daily Workflows, Project Management, Finance Management

Review Workflows

Overview

Review Workflows are periodic assessment and reflection skills that enable users to systematically evaluate their progress across different time horizons—from daily standups to yearly initiative reviews. These workflows pull data from LifeOS via MCP (Model Context Protocol) tools, present structured insights, and in some cases apply mutations to the system.

The review system follows a temporal hierarchy aligned with typical sprint and goal-setting cadences:

graph TB
    subgraph "Review Hierarchy"
        DR[Daily Review<br/>daily-plan]
        WR[Weekly Review<br/>weekly-review]
        CR[Cycle Review<br/>cycle-review]
        MR[Monthly Review<br/>monthly-review]
        IR[Initiative Review<br/>initiative-review]
    end
    
    subgraph "Specialized Reviews"
        CST[Customer Success<br/>Triage]
        BSF[Blind Spot<br/>Finder]
    end
    
    DR --> WR
    WR --> CR
    CR --> MR
    MR --> IR
    
    WR -.-> CST
    WR -.-> BSF

Each review type serves a distinct purpose in the feedback loop of planning, execution, and reflection.

Sources: skills/weekly-review/SKILL.md:1-6 Sources: skills/cycle-review/SKILL.md:1-6

Related Pages

Related topics: Review Workflows, Health Integration (Oura Ring)

Finance Management

The Finance Management module within LifeOS provides a comprehensive suite of tools for tracking, analyzing, and understanding personal finances. It integrates directly with Convex-powered backend services to deliver real-time financial insights through 126 MCP (Model Context Protocol) tools available to AI agents.

Overview

The Finance Management system consists of two primary skill workflows that enable users to:

• Monitor net worth, account balances, and financial trends
• Analyze spending patterns with detailed transaction data
• Track daily income versus spending
• Identify financial patterns and anomalies

Sources: README.md

Architecture

The Finance Management module operates within the broader LifeOS ecosystem, leveraging MCP tools to communicate with the Convex backend. All monetary values are stored in cents and converted to dollars for display.

graph TD
    A[User Request] --> B[Finance Skill Workflow]
    B --> C[MCP Tool Calls]
    C --> D[Convex Backend]
    D --> E[(Financial Data)]
    
    C --> F[get_finance_net_worth]
    C --> G[get_finance_accounts]
    C --> H[get_finance_snapshots]
    C --> I[get_finance_daily_spending]
    C --> J[get_finance_transactions]
    
    F --> K[Net Worth Dashboard]
    G --> L[Account Breakdown]
    H --> M[90-Day Trend]
    I --> N[Daily Spending Pattern]
    J --> O[Transaction History]

Sources: skills/finance-overview/SKILL.md & skills/finance-spending/SKILL.md

Available Skills

Finance Overview

The finance-overview skill provides a comprehensive financial dashboard including net worth summary, account balances, and net worth trends.

Sources: skills/finance-overview/SKILL.md

Usage:

/finance-overview

Custom Days Parameter:

/finance-overview 30

When a number is provided as an argument, it overrides the default 90-day trend period.

Finance Spending

The finance-spending skill analyzes spending patterns with daily income/spending aggregation and recent transaction details.

Sources: skills/finance-spending/SKILL.md

Usage:

/finance-spending

Custom Days Parameter:

/finance-spending 14

When a number is provided, it overrides the default 30-day analysis period.

MCP Tool Reference

The Finance Management system exposes the following MCP tools for programmatic access:

Sources: AGENTS.md

Data Models

Financial Data Types

Note: All monetary amounts are stored in cents and must be converted to dollars for display purposes.

Workflow Diagram

graph LR
    A[Start: /finance-overview] --> B[get_finance_net_worth]
    A --> C[get_finance_accounts]
    A --> D[get_finance_snapshots days=90]
    
    B --> E[Aggregate Data]
    C --> E
    D --> E
    
    E --> F[Build Dashboard]
    F --> G[Net Worth Display]
    F --> H[Assets Breakdown]
    F --> I[Liabilities Breakdown]
    F --> J[90-Day Trend]
    F --> K[Insights]

Integration Points

The Finance Management module integrates with:

• PPV (Purpose, Pillar, Vision) System - For linking financial goals to life design
• Coaching System - Financial coaching insights and action items
• Habits Tracking - Financial habit compliance
• Initiatives - Yearly financial goals tracked as initiatives

Sources: AGENTS.md

Quick Reference

All commands support numeric arguments to override default periods (e.g., /finance-overview 30 or /finance-spending 14).

Related Pages

Related topics: Habits & Accountability

Coaching System

Overview

The Coaching System is an integrated personal development framework within LifeOS that leverages multiple AI models to provide structured coaching, track action items, maintain coach memory, and help users identify blind spots. The system operates as a multi-layered coaching layer that sits atop the core LifeOS data, utilizing the MCP (Model Context Protocol) interface to access user data and apply mutations.

The coaching system serves as the "meta-layer" for personal growth, distinguishing itself from task management by focusing on identity, behavioral patterns, and sustained development rather than discrete deliverables. It coordinates with other LifeOS features like health data, habits, and finance to provide contextually-aware coaching insights.

Sources: skills/coaching-overview/SKILL.md:1-4

Architecture

The Coaching System architecture consists of three primary layers:

graph TD
    A[Coaching System] --> B[Skill Layer]
    A --> C[Memory Layer]
    A --> D[Multi-Model Layer]
    
    B --> B1[coaching-overview]
    B --> B2[coaching-action-items]
    B --> B3[coaching-session-review]
    B --> B4[coach-memory]
    
    C --> C1[Working Memory]
    C --> C2[Session Summaries]
    C --> C3[Action Items]
    
    D --> D1[llm-council]
    D --> D2[blind-spot-finder]
    
    C1 --> E[LifeOS Data]
    C2 --> E
    C3 --> E

Skill Layer

The skill layer consists of discrete workflow skills that can be invoked individually:

Sources: skills/coaching-overview/SKILL.md:1-18

Memory Layer

The memory layer maintains persistent knowledge about the user across coaching interactions:

• Core Struggles: Consistent challenges the user faces
• Values: Deeply held principles
• Triggers: Causes of spirals, procrastination, or withdrawal
• Breakthroughs: Moments of clarity or growth
• Personality: Cognitive patterns and characteristics
• Communication Style: Coaching delivery preferences
• Patterns: Recurring behavioral patterns
• Energy Sources: What energizes vs drains
• Relationships: Key relationship dynamics
• Context: Current life circumstances

Sources: skills/coach-memory/SKILL.md:8-25

Multi-Model Layer

The multi-model layer employs multiple AI models for deliberation and insight generation:

• LLM Council: Structured deliberation with peer review and synthesis
• Blind Spot Finder: Multi-model analysis to identify unknown unknowns

Sources: skills/llm-council/SKILL.md:1-5

MCP Tools

The Coaching System exposes its functionality through the following MCP tools:

Query Tools

Mutation Tools

Coaching Workflows

Dashboard Overview

The coaching-overview skill provides a consolidated dashboard by executing three parallel queries:

1. get_coaching_profiles — List all coach personas
2. get_coaching_sessions with limit=10 — Recent sessions across all coaches
3. get_coaching_action_items with status="pending" — Outstanding action items

Sources: skills/coaching-overview/SKILL.md:5-13

The dashboard presents:

• Coaches: Each coach profile with name, focus areas, and session cadence
• Recent Sessions: Grouped by coach with title, date, and status
• Pending Action Items: Count per coach with top 3 most urgent items shown
• Insights: Active coaches, overdue items, suggested next sessions

graph LR
    A[coaching-overview] --> B[get_coaching_profiles]
    A --> C[get_coaching_sessions]
    A --> D[get_coaching_action_items]
    
    B --> E[Dashboard Output]
    C --> E
    D --> E

Sources: skills/coaching-overview/SKILL.md:13-18

Session Review

The coaching-session-review skill provides in-depth review of coaching sessions:

1. Find sessions via get_coaching_sessions with limit=5
2. Resolve session ID from arguments or find most recent
3. Fetch full details via get_coaching_session

Sources: skills/coaching-session-review/SKILL.md:4-11

The session review displays:

• Session Info: Coach name, date, duration, mood at start
• Summary: AI-generated session summary
• Key Insights: All key insights from the session
• Action Items: All action items with current status
• Follow-up: Suggested next session topics based on insights and pending items

Sources: skills/coaching-session-review/SKILL.md:11-18

Action Item Management

The coaching-action-items skill manages the coaching homework cycle:

graph TD
    A[Create Action Item] --> B[Pending Status]
    B --> C[In Progress]
    C --> D{Complete?}
    D -->|Yes| E[Completed]
    D -->|No| F[Update/Extend]
    F --> C
    
    E --> G[Review in Session]
    G --> A

Action items should be:

• Specific experiments to test hypotheses
• Linked to insights from sessions or blind spot discovery
• Tracked with clear status progression

Sources: skills/blind-spot-finder/SKILL.md:27-28

Coach Memory Management

The coach-memory skill provides visibility into the AI coach's accumulated knowledge:

graph TD
    A[get_working_memory] --> B{All Sections}
    B --> C[Core Struggles]
    B --> D[Values]
    B --> E[Triggers]
    B --> F[Breakthroughs]
    B --> G[Personality]
    B --> H[Communication Style]
    B --> I[Patterns]
    B --> J[Energy Sources]
    B --> K[Relationships]
    B --> L[Context]
    
    M[Offer Update] --> N[update_working_memory]

For each section, the skill displays:

• Content of the section
• Confidence level
• Flagged sections that are empty or stale

Sources: skills/coach-memory/SKILL.md:7-25

Multi-Model Deliberation

LLM Council

The llm-council skill implements a structured multi-model deliberation process:

graph TD
    A[Start Council] --> B[Stage 1: Individual Responses]
    B --> C[Each Model Responds]
    C --> D[Stage 2: Peer Review]
    D --> E[Models Evaluate Each Other]
    E --> F[Rankings Generated]
    F --> G[Stage 3: Chairman Synthesis]
    G --> H[Final Authoritative Answer]
    
    H --> I[Optional: Save Summary]

Execution flow:

1. Stage 1 — Each council member provides an individual response
2. Stage 2 — Peer review with rankings table showing average rank per model
3. Stage 3 — Chairman synthesizes all perspectives into a final answer

Sources: skills/llm-council/SKILL.md:35-50

The council uses the zen consensus tool for deliberation. Estimated completion:

• Normal tier: 2-5 minutes
• Pro tier: 5-10 minutes

Sources: skills/llm-council/SKILL.md:33-34

Blind Spot Finder

The blind-spot-finder skill uses multi-model council to identify unknown unknowns:

graph TD
    A[Context Gathering] --> B[Working Memory]
    A --> C[Pending Action Items]
    A --> D[Habits Completion]
    A --> E[Screen Time 3 Days]
    A --> F[Sleep 14 Days]
    A --> G[Finance Net Worth]
    
    B --> H[Blind Spot Brief]
    C --> H
    D --> H
    E --> H
    F --> H
    G --> H
    
    H --> I[Multi-Model Council]
    I --> J[3-5 Blind Spots Identified]
    
    J --> K[User Chooses ONE]
    K --> L[create_coaching_action_item]
    L --> M[update_working_memory]

Sources: skills/blind-spot-finder/SKILL.md:8-28

The blind spot brief includes:

• Working memory patterns and triggers
• Say-do gaps between stated goals and actual behavior
• Incomplete action items
• Screen time vs stated priorities
• Any visible self-deceptions

Sources: skills/blind-spot-finder/SKILL.md:15-25

The council focuses on:

• Self-deceptions being maintained
• Local maxima being stuck in
• Unquestioned assumptions
• Invisible patterns from inside them
• Things systematically avoided despite stated desire

Sources: skills/blind-spot-finder/SKILL.md:25-31

Integration with LifeOS

The Coaching System integrates with other LifeOS modules to provide contextually-aware coaching:

Sources: skills/blind-spot-finder/SKILL.md:9-14

Skill Invocation

Via MCP Protocol

All coaching skills are exposed as MCP prompts, allowing any MCP client to invoke them:

/coaching-overview          # Dashboard view
/coaching-action-items      # Manage action items
/coaching-session-review    # Review session
/coach-memory               # View/manage memory
/blind-spot-finder          # Find blind spots
/llm-council                # Multi-model deliberation

Sources: AGENTS.md:1-45

Via Claude Code

claude plugin add github:starascendin/lifeos-plugin

Skills are then accessible through the Claude Code agent's skill system.

Sources: README.md:1-35

Data Models

Coaching Profile

Coaching Session

Sources: skills/coaching-session-review/SKILL.md:11-15

Working Memory Section

Sources: skills/coach-memory/SKILL.md:20-25

Coaching Action Item

Sources: skills/blind-spot-finder/SKILL.md:27-28

Best Practices

Coach Memory Maintenance

• Review memory sections regularly for accuracy
• Flag stale sections that need updating
• Update memory after breakthrough insights
• Use memory to inform coaching delivery style

Sources: skills/coach-memory/SKILL.md:24-26

Action Item Creation

• Create specific experiments, not vague goals
• Link items to identified blind spots or insights
• Use create_coaching_action_item for durable tracking
• Update working memory when genuinely new insight emerges

Sources: skills/blind-spot-finder/SKILL.md:28-29

Multi-Model Council Usage

• Be patient for deliberation results (2-10 minutes)
• Present all three stages for transparency
• Highlight individual model perspectives of value
• Offer to save insights via create_ai_convo_summary

Sources: skills/llm-council/SKILL.md:46-52

Session Review

• Always check for pending action items
• Ground insights in user's specific context
• Suggest forward-looking next steps
• Reference working memory for continuity

Sources: skills/coaching-session-review/SKILL.md:15-18

Doramagic Pitfall Log

Doramagic extracted 7 source-linked risk signals. Review them before installing or handing real data to the project.

1. Configuration risk: Configuration risk needs validation

• Severity: medium
• Finding: Configuration risk is backed by a source signal: Configuration risk needs validation. Treat it as a review item until the current version is checked.
• User impact: Users may get misleading failures or incomplete behavior unless configuration is checked carefully.
• Recommended check: Open the linked source, confirm whether it still applies to the current version, and keep the first run isolated.
• Evidence: capability.host_targets | github_repo:1156470663 | https://github.com/starascendin/lifeos-plugin | host_targets=mcp_host, claude, claude_code

2. Capability assumption: README/documentation is current enough for a first validation pass.

• Severity: medium
• Finding: README/documentation is current enough for a first validation pass.
• User impact: The project should not be treated as fully validated until this signal is reviewed.
• Recommended check: Open the linked source, confirm whether it still applies to the current version, and keep the first run isolated.
• Evidence: capability.assumptions | github_repo:1156470663 | https://github.com/starascendin/lifeos-plugin | README/documentation is current enough for a first validation pass.

3. Maintenance risk: Maintainer activity is unknown

• Severity: medium
• Finding: Maintenance risk is backed by a source signal: Maintainer activity is unknown. Treat it as a review item until the current version is checked.
• User impact: Users cannot judge support quality until recent activity, releases, and issue response are checked.
• Recommended check: Open the linked source, confirm whether it still applies to the current version, and keep the first run isolated.
• Evidence: evidence.maintainer_signals | github_repo:1156470663 | https://github.com/starascendin/lifeos-plugin | last_activity_observed missing

4. Security or permission risk: no_demo

• Severity: medium
• Finding: no_demo
• User impact: The project may affect permissions, credentials, data exposure, or host boundaries.
• Recommended check: Open the linked source, confirm whether it still applies to the current version, and keep the first run isolated.
• Evidence: downstream_validation.risk_items | github_repo:1156470663 | https://github.com/starascendin/lifeos-plugin | no_demo; severity=medium

5. Security or permission risk: no_demo

• Severity: medium
• Finding: no_demo
• User impact: The project may affect permissions, credentials, data exposure, or host boundaries.
• Recommended check: Open the linked source, confirm whether it still applies to the current version, and keep the first run isolated.
• Evidence: risks.scoring_risks | github_repo:1156470663 | https://github.com/starascendin/lifeos-plugin | no_demo; severity=medium

6. Maintenance risk: issue_or_pr_quality=unknown

• Severity: low
• Finding: issue_or_pr_quality=unknown。
• User impact: Users cannot judge support quality until recent activity, releases, and issue response are checked.
• Recommended check: Open the linked source, confirm whether it still applies to the current version, and keep the first run isolated.
• Evidence: evidence.maintainer_signals | github_repo:1156470663 | https://github.com/starascendin/lifeos-plugin | issue_or_pr_quality=unknown

7. Maintenance risk: release_recency=unknown

• Severity: low
• Finding: release_recency=unknown。
• User impact: Users cannot judge support quality until recent activity, releases, and issue response are checked.
• Recommended check: Open the linked source, confirm whether it still applies to the current version, and keep the first run isolated.
• Evidence: evidence.maintainer_signals | github_repo:1156470663 | https://github.com/starascendin/lifeos-plugin | release_recency=unknown

Community Discussion Evidence

Doramagic exposes project-level community discussion separately from official documentation. Review these links before using lifeos-plugin with real data or production workflows.

• Configuration risk needs validation - GitHub / issue