Agent skill
smart-recommendations
Stars
17
Forks
0
Install this agent skill to your Project
npx add-skill https://github.com/natea/ExoMind/tree/main/skills/smart-recommendations
SKILL.md
Smart Recommendations Skill
Purpose
AI-powered recommendation engine that analyzes patterns in your life data to provide intelligent suggestions for optimizing schedule, tasks, meetings, energy management, and goal achievement.
Core Capabilities
1. Schedule Optimization
- Best Times for Tasks: Analyzes completion history to suggest optimal time blocks
- Energy-Based Scheduling: Recommends high-focus tasks during peak energy periods
- Context Switching Reduction: Suggests task batching by context/type
- Buffer Time Insertion: Identifies where breathing room is needed
- Deadline Alignment: Recommends when to start tasks based on historical velocity
2. Task Prioritization Intelligence
- Eisenhower Matrix Auto-Assignment: Suggests urgency/importance quadrants
- Impact Scoring: Estimates task value based on goals and dependencies
- Effort Estimation: Predicts realistic time requirements from similar tasks
- Procrastination Detection: Flags repeatedly postponed tasks with suggestions
- Quick Win Identification: Highlights high-impact, low-effort opportunities
3. Meeting Optimization
- Meeting Value Assessment: Analyzes if your presence is truly needed
- Decline Suggestions: Recommends meetings to skip with reasoning
- Reschedule Opportunities: Suggests better timing for recurring meetings
- Meeting-Free Day Protection: Identifies days needing full focus time
- Preparation Time Blocking: Auto-suggests prep time before important meetings
4. Focus Time Protection
- Deep Work Block Suggestions: Recommends when to schedule uninterrupted time
- Distraction Pattern Analysis: Identifies common interruption sources
- Communication Batching: Suggests specific times for email/chat checking
- Context Preservation: Recommends how to minimize task switching
- Flow State Triggers: Identifies conditions that enable your best work
5. Energy Management
- Break Reminders: Suggests breaks based on work intensity and duration
- Energy Dip Detection: Identifies low-energy periods and suggests appropriate activities
- Recovery Time Calculation: Recommends downtime after intense periods
- Burnout Prevention: Flags unsustainable pace before it's too late
- Ultradian Rhythm Alignment: Suggests work/rest cycles matching natural patterns
6. Goal Adjustment Intelligence
- Progress Velocity Analysis: Compares actual vs planned progress
- Goal Realism Scoring: Assesses if goals are achievable given current data
- Milestone Recommendations: Suggests intermediate checkpoints
- Pivot Suggestions: Recommends when to adjust or abandon goals
- Resource Reallocation: Identifies where to shift time/energy
7. Habit Trigger Suggestions
- Implementation Intentions: Suggests "if-then" habit triggers
- Habit Stacking Opportunities: Recommends linking new habits to existing ones
- Environmental Cues: Suggests physical reminders and setup changes
- Time-Based Triggers: Identifies best times for habit execution
- Obstacle Removal: Recommends ways to reduce friction
Pattern Recognition
Data Sources Analyzed
yaml
schedule_patterns:
- Task completion times and durations
- Meeting attendance and outcomes
- Calendar density and white space
- Recurring event effectiveness
productivity_patterns:
- High-output time periods
- Task completion velocity
- Procrastination triggers
- Context switching frequency
energy_patterns:
- Daily energy fluctuations
- Break taking consistency
- Recovery time needs
- Burnout warning signs
goal_patterns:
- Progress trajectory
- Obstacle frequency
- Effort distribution
- Achievement rate
Recommendation Types
Immediate Actions
yaml
format: "RIGHT NOW: [Action]"
examples:
- "RIGHT NOW: Take a 10-minute break - you've been in flow for 2.5 hours"
- "RIGHT NOW: Decline the 3pm standup - no action items for you this sprint"
- "RIGHT NOW: Block tomorrow 9-11am for deep work - your highest energy window"
Daily Suggestions
yaml
format: "TODAY: [Suggestion]"
examples:
- "TODAY: Start the budget proposal now - historically you need 3 hours for financial docs"
- "TODAY: Move the client call to afternoon - you're 40% more patient after lunch"
- "TODAY: Batch all email responses to 4-4:30pm window"
Weekly Strategies
yaml
format: "THIS WEEK: [Strategy]"
examples:
- "THIS WEEK: Protect Tuesday for focused work - only 2 meetings currently scheduled"
- "THIS WEEK: Reduce meeting load by 3 hours - you're at 75% meeting time vs 60% target"
- "THIS WEEK: Reschedule Friday 1-on-1s to Thursday - Friday energy typically drops"
Strategic Insights
yaml
format: "INSIGHT: [Pattern]"
examples:
- "INSIGHT: You complete 3x more tasks when starting before 10am"
- "INSIGHT: Monday meetings result in 50% more action items than other days"
- "INSIGHT: Your creative work quality peaks Tuesday/Wednesday mornings"
Implementation
When to Request Recommendations
bash
# Morning routine - get daily recommendations
"What should I focus on today?"
"Any schedule optimizations for today?"
# Weekly planning - get strategic suggestions
"What are my recommendations for this week?"
"How should I adjust my schedule this week?"
# Task planning - get prioritization help
"Which tasks should I tackle first?"
"What's the best time to work on [task]?"
# Meeting review - get optimization suggestions
"Which meetings should I decline this week?"
"How can I optimize my meeting schedule?"
# Energy management - get break/rest suggestions
"When should I take breaks today?"
"Am I at risk of burnout?"
# Goal review - get adjustment recommendations
"How are my goals tracking?"
"Should I adjust any goals based on my progress?"
Recommendation Confidence Levels
yaml
HIGH_CONFIDENCE:
- Based on 10+ data points
- Pattern repeated consistently
- Strong statistical correlation
- Example: "You're 85% more productive working on reports before 10am"
MEDIUM_CONFIDENCE:
- Based on 5-9 data points
- Pattern emerging but not fully established
- Moderate correlation
- Example: "You tend to complete creative tasks faster on Tuesdays"
LOW_CONFIDENCE:
- Based on 2-4 data points
- Hypothesis for testing
- Weak but interesting correlation
- Example: "You might be more focused after morning walks"
EXPERIMENTAL:
- Based on general research/best practices
- No personal data yet
- Worth trying to gather data
- Example: "Studies show 90-minute work blocks optimize focus"
Recommendation Categories
A. Schedule Recommendations
yaml
time_blocking:
- Optimal task start times
- Best meeting windows
- Focus block placement
- Buffer time insertion
batching:
- Similar task grouping
- Context consolidation
- Communication batching
- Administrative task blocks
protection:
- Deep work preservation
- Meeting-free days
- Recovery time blocking
- No-interruption periods
B. Task Recommendations
yaml
prioritization:
- Impact vs effort scoring
- Deadline urgency ranking
- Dependency ordering
- Quick win identification
execution:
- Best time to start
- Estimated duration
- Required energy level
- Optimal context
delegation:
- Tasks to offload
- Automation opportunities
- Collaboration suggestions
- Outsourcing candidates
C. Meeting Recommendations
yaml
attendance:
- Meetings to decline
- Optional vs required
- Alternative participation (async)
- Representative delegation
scheduling:
- Better time slots
- Shorter duration opportunities
- Format changes (async, email)
- Frequency adjustments
effectiveness:
- Preparation suggestions
- Agenda improvements
- Follow-up automation
- Outcome tracking
D. Energy Recommendations
yaml
breaks:
- Optimal break timing
- Break duration suggestions
- Break activity ideas
- Movement reminders
recovery:
- End-of-day wind-down
- Weekend recharge plans
- Vacation timing
- Sabbatical consideration
prevention:
- Burnout warning signs
- Overcommitment flags
- Rest deficit alerts
- Boundary violations
Integration with Life OS
Daily Planning Integration
yaml
morning_recommendations:
- Review overnight insights
- Adjust daily plan based on suggestions
- Accept/defer/reject recommendations
- Track recommendation accuracy
real_time_suggestions:
- Pop-up notifications (optional)
- Calendar blocks with reasoning
- Task list reordering
- Energy level adjustments
Weekly Review Integration
yaml
recommendation_review:
- Accuracy assessment
- Accepted vs rejected recommendations
- Impact of followed suggestions
- Pattern refinement feedback
learning_loop:
- Update recommendation models
- Incorporate new patterns
- Adjust confidence levels
- Refine algorithms
Goal Review Integration
yaml
goal_recommendations:
- Progress trajectory analysis
- Milestone adjustment suggestions
- Resource reallocation ideas
- Goal abandonment signals
quarterly_insights:
- Major pattern discoveries
- Successful recommendation types
- Areas for improvement
- New recommendation categories
Privacy and Control
Recommendation Settings
yaml
frequency:
- real_time: Enable/disable live suggestions
- daily: Morning recommendation summary
- weekly: Sunday night strategy brief
- monthly: Pattern insight report
categories:
- schedule_optimization: true/false
- task_prioritization: true/false
- meeting_management: true/false
- energy_management: true/false
- goal_adjustments: true/false
notification_style:
- subtle: Show in review only
- moderate: Daily digest
- active: Real-time suggestions
- aggressive: Proactive interruptions
Data Privacy
- All analysis happens locally
- No external AI model calls for sensitive data
- User can delete recommendation history
- Opt-in for specific data sources
- Export recommendation logs for review
Recommendation Quality Metrics
Track Effectiveness
yaml
metrics:
acceptance_rate:
- Percentage of recommendations followed
- By category and confidence level
outcome_tracking:
- Did following recommendation improve results?
- Task completion, energy levels, goal progress
false_positives:
- Recommendations that seemed good but weren't
- Reasons for rejection
missed_opportunities:
- Patterns you noticed that weren't recommended
- Gaps in recommendation coverage
Continuous Improvement
yaml
learning_mechanisms:
- User feedback (thumbs up/down)
- Outcome correlation analysis
- A/B testing different recommendation types
- Pattern discovery algorithms
- Confidence level calibration
Example Recommendation Scenarios
Scenario 1: Morning Energy Optimization
DATA PATTERN:
- Task completion 40% higher before 10am
- Creative work quality highest 8-10am
- Deep work sessions most successful early morning
- Energy dips significantly after lunch
RECOMMENDATIONS:
HIGH CONFIDENCE:
✓ Schedule all creative work between 8-10am
✓ Block 8-11am for deep work (no meetings)
✓ Move recurring team standup to 2pm
✓ Batch email responses to afternoon
MEDIUM CONFIDENCE:
• Consider earlier wake time (7am vs 8am)
• Experiment with morning exercise before work
• Limit coffee to one cup before focus blocks
EXPERIMENTAL:
? Try cold shower before creative work
? Test 90-minute focus blocks vs 60-minute
Scenario 2: Meeting Overload
DATA PATTERN:
- 28 hours of meetings last week (70% of work time)
- Only 4 meetings had actionable outcomes
- 60% of meeting time spent listening to updates
- Productivity 3x lower on heavy meeting days
RECOMMENDATIONS:
HIGH CONFIDENCE:
✓ DECLINE: Daily standups Mon/Wed/Fri (contribute async)
✓ DECLINE: Project review meeting (get recording + notes)
✓ SHORTEN: Weekly 1-on-1s from 60min to 30min
✓ RESCHEDULE: Client calls to Tuesday/Thursday (better energy)
MEDIUM CONFIDENCE:
• Propose alternating weekly all-hands attendance
• Batch all internal meetings to Monday/Friday
• Request agenda + materials 24hrs before joining
ACTION PLAN:
1. This week: Decline 3 lowest-value recurring meetings
2. Next week: Test async participation for standups
3. Track: Productivity gain from reclaimed time
Scenario 3: Procrastination Pattern
DATA PATTERN:
- "Write Q4 proposal" rescheduled 7 times
- All writing tasks delayed until deadline pressure
- 3x more time spent on writing when under pressure
- Better quality when given 2-3 day buffer
RECOMMENDATIONS:
HIGH CONFIDENCE:
✓ START NOW: Block tomorrow 9-11am for proposal draft
✓ Break into smaller tasks (outline, draft sections, review)
✓ Set fake deadline 3 days before real deadline
✓ Schedule accountability check-in with peer
MEDIUM CONFIDENCE:
• Work on proposal in 25-minute Pomodoro sessions
• Change environment (coffee shop, library)
• Reward completion with something you enjoy
ROOT CAUSE ANALYSIS:
- Task feels overwhelming (need smaller chunks)
- Unclear requirements (need to clarify scope)
- Perfectionism (need to embrace "good enough" draft)
- Low interest (consider if this is right work)
PREVENTION:
→ For future writing tasks, start 5 days before deadline
→ Create proposal template to reduce cognitive load
→ Partner with colleague who enjoys writing
Scenario 4: Burnout Prevention
DATA PATTERN:
- Working 60+ hours/week for 4 consecutive weeks
- Zero days completely off in 3 weeks
- Sleep quality decreased 30%
- Completion rate dropped despite more hours
- Irritability and decision fatigue increasing
RECOMMENDATIONS:
HIGH CONFIDENCE:
✓ URGENT: Take full day off this weekend (no laptop)
✓ Reduce work hours to 45/week for next 2 weeks
✓ Decline all new commitments until recovery
✓ Schedule 7+ hours sleep nightly (track with alarm)
IMMEDIATE ACTIONS:
✓ Cancel tomorrow's optional meetings
✓ Delegate project X to team member
✓ Push deadline for proposal Y by 1 week
✓ Block Friday afternoon as "recovery time"
STRATEGIC CHANGES:
• Audit all commitments and eliminate 20%
• Establish "shutdown ritual" at 6pm daily
• Schedule weekly sabbath (full day off)
• Set up burnout early warning system
WARNING SIGNS TO MONITOR:
- Continued poor sleep
- Rising irritability
- Decreased task completion
- Physical symptoms (headaches, tension)
- Loss of interest in previously enjoyed activities
If burnout continues despite interventions:
→ Consider taking 1-2 week vacation
→ Discuss workload with manager
→ Evaluate if role/job is sustainable
Advanced Features
Predictive Recommendations
yaml
anticipatory_suggestions:
- "Big deadline next month - start blocking focus time now"
- "Travel next week - batch meetings before/after"
- "Typical Q4 crunch coming - frontload important projects"
- "Energy usually dips in winter - adjust expectations"
Comparative Analysis
yaml
compare_to:
- Your past self (last quarter, last year)
- Anonymized peer patterns (similar roles)
- Research-based best practices
- Your stated ideal schedule
insights:
- "You're having 40% more meetings than last quarter"
- "Your focus time is 2x higher than typical for role"
- "Sleep pattern optimal compared to circadian research"
Scenario Testing
yaml
what_if_analysis:
- "What if I decline recurring meeting X?"
- "What if I shift wake time to 6am?"
- "What if I batch all meetings to 2 days/week?"
- "What if I work 4-day weeks?"
estimated_impact:
- Reclaimed time calculation
- Energy level prediction
- Productivity change estimate
- Goal progress trajectory
Success Metrics
Recommendation Adoption
- % of recommendations viewed
- % of recommendations accepted
- Time to act on recommendations
- Category-specific adoption rates
Impact Measures
- Task completion rate change
- Goal progress velocity change
- Meeting time reduction
- Focus time increase
- Energy level improvement
- Burnout risk reduction
System Learning
- Recommendation accuracy over time
- Pattern discovery rate
- False positive reduction
- User satisfaction scores
- Recommendation diversity
Getting Started
- Enable Recommendations: Turn on recommendation engine in settings
- Configure Preferences: Set notification level and categories
- Build Data Foundation: 2-4 weeks of tracking for initial patterns
- Review First Recommendations: Start with daily digest
- Provide Feedback: Rate recommendations to improve accuracy
- Iterate: Adjust settings based on what's helpful vs noisy
Related Skills
- Daily Planning - Apply recommendations during planning
- Weekly Review - Assess recommendation effectiveness
- Tracking Habits - Get habit optimization suggestions
- Processing Inbox - Prioritization recommendations
- Goal Setting - Goal adjustment recommendations
Smart recommendations transform data into actionable intelligence. Let patterns guide your optimization.
Didn't find tool you were looking for?