executive（executive 搜索）

Executive Skill

Analyze productivity, prioritize work, detect patterns, and generate strategic recommendations.

When to Use

• User asks "What's most important?" or "What should I focus on?"
• User asks "How am I doing?" or "How productive have I been?"
• User wants productivity analysis (daily, weekly, monthly)
• Generating weekly or monthly strategic reviews
• Detecting bottlenecks, stagnation, or work distribution issues
• Recommending focus areas or schedule adjustments
• User asks about patterns or trends in their work
• User wants goal progress assessment or risk analysis

Database Location

SECRETARY_DB_PATH="$HOME/.claude/secretary/secretary.db"

Work Prioritization

Priority Scoring

Score each commitment by multiple weighted factors:

SELECT
    c.id, c.title, c.due_date, c.priority, c.project, c.stakeholder,
    -- Urgency (0-100)
    CASE
        WHEN c.due_date < date('now') THEN 100
        WHEN c.due_date = date('now') THEN 80
        WHEN c.due_date <= date('now', '+2 days') THEN 60
        WHEN c.due_date <= date('now', '+7 days') THEN 40
        ELSE 20
    END as urgency,
    -- Priority weight (0-40)
    CASE c.priority
        WHEN 'critical' THEN 40
        WHEN 'high' THEN 30
        WHEN 'medium' THEN 20
        ELSE 10
    END as priority_score,
    -- Stakeholder factor (0-20)
    CASE WHEN c.stakeholder IS NOT NULL THEN 20 ELSE 0 END as stakeholder_score,
    -- Deferral penalty (-10 per deferral)
    -10 * c.deferred_count as deferral_penalty
FROM commitments c
WHERE c.status IN ('pending', 'in_progress')
ORDER BY (urgency + priority_score + stakeholder_score + deferral_penalty) DESC;

Eisenhower Matrix

Categorize work:

URGENT + IMPORTANT     -> Do First
NOT URGENT + IMPORTANT -> Schedule
URGENT + NOT IMPORTANT -> Delegate
NOT URGENT + NOT IMPORTANT -> Eliminate

Productivity Analysis

Session Metrics

-- Daily summary (last 30 days)
SELECT
    date(started_at) as date,
    COUNT(*) as sessions,
    SUM(duration_seconds) / 3600.0 as hours,
    AVG(duration_seconds) / 60.0 as avg_minutes,
    COUNT(DISTINCT project) as projects
FROM sessions
WHERE started_at >= datetime('now', '-30 days') AND status = 'completed'
GROUP BY date(started_at)
ORDER BY date DESC;

Completion Rates

-- Weekly completion rate (last 90 days)
SELECT
    strftime('%Y-%W', created_at) as week,
    COUNT(*) as created,
    SUM(CASE WHEN status = 'completed' THEN 1 ELSE 0 END) as completed,
    ROUND(100.0 * SUM(CASE WHEN status = 'completed' THEN 1 ELSE 0 END) / COUNT(*), 1) as rate,
    SUM(CASE WHEN status = 'deferred' THEN 1 ELSE 0 END) as deferred,
    SUM(CASE WHEN status = 'canceled' THEN 1 ELSE 0 END) as canceled
FROM commitments
WHERE created_at >= datetime('now', '-90 days')
GROUP BY week
ORDER BY week DESC;

Time Distribution

-- Where is time going? (last 7 days)
SELECT
    project,
    SUM(duration_seconds) / 3600.0 as hours,
    ROUND(100.0 * SUM(duration_seconds) / (
        SELECT SUM(duration_seconds) FROM sessions
        WHERE started_at >= datetime('now', '-7 days') AND status = 'completed'
    ), 1) as percentage,
    COUNT(*) as sessions
FROM sessions
WHERE started_at >= datetime('now', '-7 days') AND status = 'completed'
GROUP BY project
ORDER BY hours DESC;

Decision Velocity

-- How often decisions are being made
SELECT
    strftime('%Y-%W', created_at) as week,
    COUNT(*) as total,
    SUM(CASE WHEN category = 'architecture' THEN 1 ELSE 0 END) as architecture,
    SUM(CASE WHEN category = 'technology' THEN 1 ELSE 0 END) as technology,
    SUM(CASE WHEN category = 'process' THEN 1 ELSE 0 END) as process,
    SUM(CASE WHEN category = 'design' THEN 1 ELSE 0 END) as design
FROM decisions
WHERE created_at >= datetime('now', '-90 days')
GROUP BY week
ORDER BY week DESC;

Goal Tracking

Progress Velocity

SELECT
    g.id, g.title, g.goal_type, g.progress_percentage,
    g.target_date, g.target_value, g.current_value, g.target_unit,
    -- Days remaining
    ROUND(julianday(g.target_date) - julianday('now'), 0) as days_remaining,
    -- Required daily progress
    ROUND((100 - g.progress_percentage) / NULLIF(julianday(g.target_date) - julianday('now'), 0), 2) as required_daily,
    -- Risk status
    CASE
        WHEN g.target_date < date('now') AND g.progress_percentage < 100 THEN 'overdue'
        WHEN (100 - g.progress_percentage) / NULLIF(julianday(g.target_date) - julianday('now'), 0) > 5 THEN 'at_risk'
        WHEN (100 - g.progress_percentage) / NULLIF(julianday(g.target_date) - julianday('now'), 0) > 2 THEN 'needs_attention'
        ELSE 'on_track'
    END as risk_status
FROM goals g
WHERE g.status = 'active' AND g.target_date IS NOT NULL
ORDER BY days_remaining ASC;

Bottleneck Detection

Long-Pending Items

SELECT id, title, priority, project,
    ROUND(julianday('now') - julianday(created_at), 0) as days_pending
FROM commitments
WHERE status = 'pending'
ORDER BY days_pending DESC
LIMIT 10;

Frequently Deferred

SELECT id, title, deferred_count, priority, project
FROM commitments
WHERE deferred_count >= 2
  AND status NOT IN ('completed', 'canceled')
ORDER BY deferred_count DESC;

Blocked Work

-- Items that have been in_progress for a long time
SELECT id, title, priority, project,
    ROUND(julianday('now') - julianday(updated_at), 0) as days_stale
FROM commitments
WHERE status = 'in_progress'
  AND updated_at < datetime('now', '-7 days')
ORDER BY days_stale DESC;

Pattern Integration

Use detected patterns to inform recommendations:

SELECT id, title, pattern_type, category, confidence,
    evidence_count, recommendations, frequency
FROM patterns
WHERE status = 'active' AND confidence >= 0.6
ORDER BY confidence DESC, evidence_count DESC;

Apply insights:

• Productivity timing - Schedule important work during peak hours
• Work style - Recommend process improvements
• Completion patterns - Adjust estimation and scheduling
• Context switching - Recommend focus blocks
• Deferral patterns - Flag items that need attention or elimination

Output Guidelines

Priority Report

# Work Priorities

## Do Today
1. **[C-0001] Fix auth bug** - Overdue, critical
   - Stakeholder: Product team
   - Estimated: 2 hours

2. **[C-0003] Review PR** - Due today
   - Impact: Team velocity

## Schedule This Week
3. **[G-0002] API integration** - Goal at risk (60%)
   - Required: +8%/day
4. **[C-0005] Update docs** - Low urgency, high value

## Consider Dropping
- [C-0010] Research task - 3x deferred, low impact

Strategic Review

# Weekly Review

## Metrics
| Metric | This Week | Trend |
|--------|-----------|-------|
| Sessions | 18 | +20% |
| Hours | 24h | +15% |
| Completed | 11 | +37% |
| Decisions | 5 | stable |

## Concerns
1. Carryover growing - 8 items >2 weeks
2. Goal G-0003 at risk - needs 20%/week
3. Context switching - 5 projects in 7 days

## Recommendations
1. Clear backlog tomorrow (2h)
2. Documentation sprint Wednesday
3. Weekly planning Friday

## Wins
- Morning productivity pattern continues
- Test coverage improving steadily

Review Types

Weekly Focus

• Top 3 priorities for the week
• Carryover items analysis
• Goal trajectory check
• Time allocation review
• Pattern-based scheduling suggestions

Monthly Strategy

• Progress on major objectives
• Decision patterns analysis (categories, reversal rate)
• Productivity trends (velocity, efficiency)
• Knowledge graph growth
• Recommendation adjustments

Recommendation Categories

1. Immediate Actions - What to do right now
2. Scheduling - When to do what, based on data
3. Process Changes - How to improve workflows
4. Risk Mitigation - What to watch for
5. Celebration - What went well (motivation matters)

Error Handling

• Handle empty datasets gracefully ("Not enough data yet for analysis")
• Provide default recommendations when data is sparse
• Note when analysis is based on limited data
• Degrade gracefully if specific tables have no rows

Related Commands

• /secretary:review - Generate weekly or monthly review
• /secretary:patterns - View detected work patterns
• /secretary:goals - Manage and track goals
• /secretary:status - Full dashboard with metrics
• /secretary:priorities - Ranked priority list