claude-home/productivity/CONTEXT.md
Cal Corum c8dcf2b5ee CLAUDE: Add productivity tools with n8n workflow automation
- Add CONTEXT.md with ADHD-optimized task management patterns
- Add troubleshooting guide for productivity tools
- Add n8n workflow documentation including Ko-fi integration
- Document n8n at LXC 210 (10.10.0.210)

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
2025-12-07 00:48:28 -06:00

9.3 KiB

Productivity Tools - Technology Context

Overview

ADHD-optimized task management system designed to break the context-switching spiral. The system uses a persistent terminal dashboard with zero-friction brain dump capture to help maintain focus while capturing interrupting thoughts.

Architecture Patterns

Single-Focus with Brain Dump Pattern

Pattern: One visible task at a time with instant thought capture

  • Current Focus: Single task display with time tracking
  • Brain Dump: Zero-friction capture without context switching
  • Auto-Progression: System-managed task transitions
  • Priority Queue: Automatic high→medium→low ordering

When to Use:

  • ADHD workflow challenges
  • Frequent context switching problems
  • "Meant to do" task accumulation
  • Need for always-visible task tracking
  • Terminal-based work environments

Design Principles

  1. Minimal Friction: Two-second task capture without switching
  2. Always Visible: Persistent terminal dashboard prevents forgetting
  3. Single Focus: One task display reduces overwhelm
  4. Auto-Progression: System picks next task, reducing decision fatigue
  5. Time Awareness: Duration tracking without judgment

Core Components

Task Manager Core

Purpose: Central task management with priority queuing and state management Location: ~/.claude/tools/task-manager/task_manager.py

Key Features:

  • Priority-based task ordering (High/Medium/Low)
  • Task states: brain_dump, current, completed, snoozed
  • Auto-progression on task completion
  • Snooze functionality with time-based wake-up
  • JSON-based persistence

Data Model:

class Task:
    id: str                      # Timestamp-based unique ID
    description: str             # Task description
    priority: str                # high, medium, low
    status: str                  # brain_dump, current, completed, snoozed
    created_at: str             # ISO timestamp
    started_at: Optional[str]   # When task became current
    completed_at: Optional[str] # Completion timestamp
    snoozed_until: Optional[str] # Wake-up time for snoozed tasks

Terminal Dashboard

Purpose: Persistent visual display of current focus and brain dump queue Location: ~/.claude/tools/task-manager/dashboard.py

Display Components:

  • Current focus with elapsed time
  • Brain dump queue (up to 8 tasks visible)
  • Completion statistics
  • Command reference

Update Behavior:

  • Auto-refreshes every 1 second
  • Reads from JSON file (detects CLI changes)
  • Uses rich library for clean TUI rendering

CLI Interface

Purpose: Zero-friction command-line task operations Location: ~/.claude/tools/task-manager/cli.py

Core Commands:

  • task dump "description" - Brain dump capture (most important!)
  • task done - Complete current, auto-load next
  • task next - Skip to next task
  • task snooze [hours] - Snooze current task
  • task add "description" --now - Add and make current
  • task - Show current status

Priority Flags:

  • --high - High priority (🔴)
  • --medium - Medium priority (🟡) [default]
  • --low - Low priority (🟢)

Implementation Patterns

The Context-Switch Spiral Solution

Problem Pattern:

Working on Task A → Remember Task B → Switch to Task B →
Remember Task C → Task A forgotten → Task B becomes "meant to do" → Loop

Solution Pattern:

Working on Task A → Remember Task B → `task dump "Task B"` →
Continue Task A → Complete with `task done` → System auto-loads Task B

Brain Dump Workflow

  1. Capture: task dump "thought" (2 seconds, no context switch)
  2. Visible: Task appears in brain dump queue on dashboard
  3. Continue: Keep working on current focus
  4. Auto-Load: Task surfaces when current task completes

Priority Management

Automatic Ordering:

  1. High priority tasks (🔴) surface first
  2. Medium priority tasks (🟡) second
  3. Low priority tasks (🟢) last
  4. Within priority level, oldest tasks first

Use Cases:

  • --high: Urgent, time-sensitive, or blocking tasks
  • --medium: Normal daily work (default)
  • --low: "Nice to have", research, exploration

Time Tracking Pattern

Purpose: Awareness without judgment

  • Shows elapsed time on current task
  • No alerts or pressure
  • Helps identify time sinks naturally
  • User decides when to move on

Common Workflows

Morning Startup

# Brain dump everything you need to do
task dump "Review PRs" --high
task dump "Update documentation"
task dump "Research new library" --low

# Load first task (highest priority)
task next

# Start dashboard in persistent pane
task-dashboard

During Work Day

# Working on current task... suddenly remember something
task dump "email Staci about dinner"

# Thought is captured, visible in brain dump, keep working

# Finish current task
task done
# → Auto-loads next task from queue

Handling Interruptions

# Urgent task comes up
task add "Fix production bug" --now --high
# → Immediately becomes current focus
# → Previous task moved back to brain dump

# Can't work on current task right now
task snooze 2
# → Snoozed for 2 hours
# → Next task auto-loaded

End of Day

# Check what's left
task

# Everything in brain dump waits for tomorrow
# Completed tasks tracked in stats

Best Practices

Dashboard Setup

Persistent Pane Strategy:

  • Use tmux/screen split pane for dashboard
  • Keep visible on secondary monitor
  • Always-visible = can't forget to check
  • Alternative: Dedicated terminal window

Terminal Multiplexer Example (tmux):

# Split terminal horizontally
tmux split-window -h

# Run dashboard in right pane
task-dashboard

# Work in left pane

Brain Dump Discipline

Key Principles:

  • Capture thoughts immediately when they appear
  • Don't evaluate or organize - just dump
  • Trust the system to surface them later
  • Priority can be added later if wrong

Anti-Patterns to Avoid:

  • Stopping work to add detailed task information
  • Trying to organize tasks while capturing
  • Skipping capture because task seems "small"
  • Context switching to "quickly do" the task

Priority Usage

High Priority - Use sparingly:

  • Urgent deadlines (today/this week)
  • Blocking other people's work
  • Production issues
  • Time-sensitive opportunities

Medium Priority - Default for most work:

  • Regular daily tasks
  • Normal project work
  • Scheduled activities

Low Priority - Background/optional:

  • Research and exploration
  • "Someday/maybe" items
  • Nice-to-have improvements

System Maintenance

Daily:

  • Complete tasks as you work
  • Brain dump new thoughts immediately
  • Trust auto-progression

Weekly:

  • Review low-priority items
  • Clean up stale tasks
  • Adjust priorities if needed

Integration Patterns

PATH Integration

Location: ~/.bashrc

export PATH="$HOME/.claude/tools/task-manager:$PATH"

Benefits:

  • task and task-dashboard available globally
  • No full paths needed
  • Muscle memory development

Data Storage

Location: ~/.claude/tools/task-manager/tasks.json Format: JSON with task array and metadata

Backup Strategy:

  • File-based storage enables easy backup
  • Can sync via Dropbox/Google Drive if desired
  • Simple JSON enables scripting/reporting

Future Integration Points

Planned Enhancements:

  • Notion MCP sync (Cal has integration available)
  • Discord notifications on task completion
  • Scheduled Jarvis check-ins (periodic reminders)
  • Voice capture integration
  • Pomodoro timer integration
  • Analytics and reporting

Technical Details

File Structure

~/.claude/tools/task-manager/
├── task_manager.py      # Core task management logic
├── dashboard.py         # TUI dashboard (persistent display)
├── cli.py              # Command-line interface
├── task                # Shell wrapper for CLI
├── task-dashboard      # Shell wrapper for dashboard
├── tasks.json          # Data storage (auto-created)
└── README.md           # User documentation

Dependencies

  • Python 3.7+
  • rich library for TUI rendering
  • No other external dependencies

Installation

Already Installed - Cal's system:

  • Scripts: ~/.claude/tools/task-manager/
  • PATH: Added to ~/.bashrc
  • Data: ~/.claude/tools/task-manager/tasks.json

ADHD-Specific Optimizations

Why Traditional Systems Fail

  1. Hidden in apps - Out of sight, out of mind
  2. Too much organization - Analysis paralysis
  3. Too many features - Overwhelm and abandonment
  4. No immediate feedback - Delayed dopamine
  5. Context switching required - Defeats purpose

How This System Addresses ADHD

  1. Always visible - Persistent terminal display
  2. Minimal organization - Just dump and go
  3. Minimal features - Only what's needed
  4. Immediate feedback - Instant completion confirmation
  5. Zero context switch - Capture without abandoning work

The Psychology

Reduced Anxiety: Thoughts are captured and visible Maintained Focus: No need to switch away from current work Reduced Overwhelm: Single focus display, not full list Dopamine Feedback: Immediate confirmation on completion Decision Reduction: System picks next task automatically

This productivity system is optimized specifically for ADHD workflows and terminal-based work environments, providing the minimal viable structure needed to maintain focus while capturing interrupting thoughts.