resilience-dashboard

name	resilience-dashboard
description	Generate a comprehensive resilience status report using all available MCP tools. Aggregates unified critical index, burnout Rt, early warnings, utilization, and defense levels into a single actionable dashboard.
model_tier	sonnet
parallel_hints	{"can_parallel_with":["schedule-validator","acgme-compliance"],"must_serialize_with":["safe-schedule-generation"],"preferred_batch_size":8}
context_hints	{"max_file_context":50,"compression_level":1,"requires_git_context":false,"requires_db_context":true}
escalation_triggers	[{"pattern":"RED|BLACK","reason":"Critical status requires immediate human attention"},{"pattern":"sacrifice.*hierarchy","reason":"Sacrifice hierarchy activation requires human approval"},{"keyword":["emergency","crisis","N-2 failure"],"reason":"Emergency situations require human decision-making"}]

Resilience Dashboard

Purpose: Generate a comprehensive real-time resilience status report by orchestrating multiple MCP resilience tools in parallel Created: 2025-12-28 Trigger: /resilience-dashboard command

---

When to Use

• Morning standup to get program health overview
• Before major scheduling decisions (block generation, swaps)
• After emergency events (absence, deployment notification)
• Weekly/monthly resilience health checks
• When program coordinator asks "How healthy is our schedule?"
• Investigating unexplained coverage gaps or faculty stress
• Post-incident retrospective analysis

---

Required Actions

When this skill is invoked, Claude MUST:

1. Call MCP tools in parallel to gather resilience metrics:

   Parallel Batch 1 - Core Metrics:

   - get_unified_critical_index_tool(include_details=True, top_n=5)
   - calculate_burnout_rt_tool(burned_out_provider_ids=[], time_window_days=28)
   - check_utilization_threshold_tool(available_faculty, required_blocks)
   - get_defense_level_tool(coverage_rate)
   

   Parallel Batch 2 - Early Warning:

   - detect_burnout_precursors_tool (for any flagged residents)
   - run_spc_analysis_tool (if weekly hours data available)
   - calculate_fire_danger_index_tool (for high-risk residents)
   

   Parallel Batch 3 - Contingency:

   - run_contingency_analysis_resilience_tool(analyze_n1=True, analyze_n2=True)
   

2. Determine overall status using the severity hierarchy:

   • GREEN: All metrics healthy, no interventions needed
   • YELLOW: Minor concerns, monitoring recommended
   • ORANGE: Elevated risk, action recommended within 24 hours
   • RED: Critical issues, immediate intervention required
   • BLACK: System in crisis mode, emergency protocols active

3. Aggregate and synthesize the results into a unified dashboard

4. Highlight critical issues requiring immediate attention

5. Provide trend analysis if historical data is available

6. Recommend prioritized actions based on highest-impact interventions

---

Output Format

Generate the dashboard in this exact markdown format:

# Resilience Dashboard

**Generated:** [ISO timestamp]
**Period:** [date range analyzed]

## Overall Status: [GREEN/YELLOW/ORANGE/RED/BLACK]

[1-2 sentence executive summary]

---

### Critical Index: [0-100]

| Domain | Score | Status |
|--------|-------|--------|
| Contingency (N-1/N-2) | [score] | [status] |
| Epidemiology (Burnout Spread) | [score] | [status] |
| Hub Analysis (Network) | [score] | [status] |

**Top Priority Faculty:**
1. [faculty_id]: [risk pattern] - [recommended intervention]
2. ...

**Contributing Factors:**
- [factor 1]
- [factor 2]

---

### Burnout Epidemic Status

| Metric | Value | Interpretation |
|--------|-------|----------------|
| Rt (Reproduction Number) | [value] | [spreading/controlled/declining] |
| Status | [status] | [description] |
| Intervention Level | [level] | [none/monitoring/moderate/aggressive/emergency] |

**Superspreaders Identified:** [count]
**Herd Immunity Threshold:** [%]

**Interventions:**
- [intervention 1]
- [intervention 2]

---

### Early Warning Signals

| Signal Type | Alerts | Severity |
|-------------|--------|----------|
| Seismic Precursors | [count] | [severity] |
| SPC Violations | [count] | [severity] |
| Fire Danger Index | [avg FWI] | [danger class] |

**Residents Requiring Attention:**
- [resident_id]: [signal type] - [recommended action]

---

### Capacity Status

| Metric | Value | Threshold | Status |
|--------|-------|-----------|--------|
| Utilization Rate | [%] | 80% | [status] |
| Defense Level | [level] | Prevention | [status] |
| Buffer Remaining | [%] | 20% | [status] |

**Wait Time Multiplier:** [value]x

---

### Contingency Analysis

| Analysis | Pass/Fail | Vulnerabilities |
|----------|-----------|-----------------|
| N-1 (Single Absence) | [pass/fail] | [count] critical |
| N-2 (Dual Absence) | [pass/fail] | [count] fatal pairs |

**Most Critical Faculty:** [list]

**Phase Transition Risk:** [low/medium/high/critical]

---

### Recommended Actions

**Immediate (within 24 hours):**
1. [action with specific target]
2. ...

**Short-term (within 1 week):**
 [action with specific target]
 ...


 [action with specific target]
 ...

---

---

Tool Orchestration Details

MCP Tools Called

Tool	Purpose	Required Inputs	Criticality
get_unified_critical_index_tool	Multi-factor risk aggregation	None (auto-fetches)	HIGH
calculate_burnout_rt_tool	Burnout epidemic modeling	burned_out_provider_ids	HIGH
detect_burnout_precursors_tool	Seismic STA/LTA detection	resident_id, signal_type, time_series	MEDIUM
run_spc_analysis_tool	Western Electric rules	resident_id, weekly_hours	MEDIUM
calculate_fire_danger_index_tool	CFFDRS burnout prediction	resident_id, metrics	MEDIUM
check_utilization_threshold_tool	80% queuing threshold	available_faculty, required_blocks	HIGH
get_defense_level_tool	Defense-in-depth status	coverage_rate	HIGH
run_contingency_analysis_resilience_tool	N-1/N-2 vulnerability	Analysis flags	HIGH

Parallel Execution Strategy

Time ─────────────────────────────────────────────────────►

Phase 1 (Core):
├── get_unified_critical_index_tool ─────────┐
├── calculate_burnout_rt_tool ───────────────┤
├── check_utilization_threshold_tool ────────┤  All parallel
└── get_defense_level_tool ──────────────────┘

Phase 2 (Early Warning) - after Phase 1 if flagged residents:
├── detect_burnout_precursors_tool ──────────┐
├── run_spc_analysis_tool ───────────────────┤  Parallel
└── calculate_fire_danger_index_tool ────────┘

Phase 3 (Contingency) - runs in parallel with Phase 2:
└── run_contingency_analysis_resilience_tool ─┘

Phase 4: Aggregate ──► Format ──► Output

Status Determination Logic

def determine_overall_status(results):
    """
    Determine overall status from aggregated tool results.

    Severity precedence: BLACK > RED > ORANGE > YELLOW > GREEN
    """
    # Any emergency-level result = BLACK
    if any(r.severity == "emergency" for r in results):
        return "BLACK"

    # Any critical-level result = RED
    if any(r.severity == "critical" for r in results):
        return "RED"

    # Any elevated/warning with high impact = ORANGE
    if (results.unified_index.overall_index >= 60 or
        results.burnout_rt.rt >= 1.5 or
        results.utilization.utilization_rate >= 0.85):
        return "ORANGE"

    # Any warning-level result = YELLOW
    if any(r.severity in ["warning", "elevated"] for r in results):
        return "YELLOW"

    return "GREEN"

---

Examples

Example 1: Healthy Program

User: /resilience-dashboard

Output:

# Resilience Dashboard

**Generated:** 2025-12-28T10:30:00Z
**Period:** 2025-12-01 to 2025-12-28

## Overall Status: GREEN

All resilience metrics are within healthy thresholds. No immediate interventions required.

### Critical Index: 28

| Domain | Score | Status |
|--------|-------|--------|
| Contingency (N-1/N-2) | 0.25 | Healthy |
| Epidemiology (Burnout Spread) | 0.20 | Healthy |
| Hub Analysis (Network) | 0.30 | Healthy |

... [rest of dashboard with green metrics]

### Recommended Actions

**Immediate (within 24 hours):**
1. None required

**Short-term (within 1 week):**
1. Continue routine monitoring

*Next recommended refresh: 7 days*

Example 2: Elevated Risk

User: /resilience-dashboard

Output:

# Resilience Dashboard

**Generated:** 2025-12-28T10:30:00Z
**Period:** 2025-12-01 to 2025-12-28

## Overall Status: ORANGE

Elevated risk detected in contingency analysis. Two faculty members identified as single points of failure. Recommend cross-training within 24 hours.

### Critical Index: 58

| Domain | Score | Status |
|--------|-------|--------|
| Contingency (N-1/N-2) | 0.72 | CRITICAL |
| Epidemiology (Burnout Spread) | 0.35 | Warning |
| Hub Analysis (Network) | 0.55 | Elevated |

**Top Priority Faculty:**
1. FAC-007: STRUCTURAL_BURNOUT - Immediate workload reduction
2. FAC-012: INFLUENTIAL_HUB - Cross-training priority

... [rest of dashboard]

### Recommended Actions

**Immediate (within 24 hours):**
1. Begin cross-training for FAC-007 backup coverage
2. Review FAC-012 schedule for workload distribution

*Next recommended refresh: 24 hours*

Example 3: Crisis Mode

User: /resilience-dashboard (during deployment event)

Output:

# Resilience Dashboard

**Generated:** 2025-12-28T10:30:00Z
**Period:** 2025-12-01 to 2025-12-28

## Overall Status: RED

CRITICAL: Multiple faculty absences due to deployment have triggered N-2 failure state.
Defense level escalated to CONTAINMENT. Sacrifice hierarchy activation recommended.

### Critical Index: 82

... [dashboard with critical metrics]

### Recommended Actions

**Immediate (within 24 hours):**
1. ESCALATE: Request human approval for sacrifice hierarchy activation
2. Activate static fallback schedule "deployment"
3. Contact backup pool for emergency coverage
4. Generate SITREP for program director

*Next recommended refresh: 4 hours*

---

Escalation Rules

Escalate to human when:

1. Overall status is RED or BLACK
2. N-2 analysis shows fatal faculty pairs
3. Burnout Rt >= 2.0 (rapid spread)
4. Defense level reaches CONTAINMENT or EMERGENCY
5. Sacrifice hierarchy execution is recommended
6. Multiple universal-critical faculty identified
7. Utilization exceeds 90%

Can handle automatically:

1. GREEN/YELLOW status reporting
2. Data aggregation and formatting
3. Trend analysis from historical data
4. Routine monitoring recommendations
5. Identification of at-risk individuals
6. Prioritization of action items

---

Error Handling

If MCP tools fail, the dashboard should:

1. Report partial results with clear indication of missing data
2. Note which tools failed and potential impact
3. Provide fallback recommendations based on available data
4. Suggest retry if transient failures suspected

Example partial output:

## Overall Status: YELLOW (Partial Data)

**Warning:** Some metrics unavailable due to tool failures:
- get_unified_critical_index_tool: Timeout (retrying...)
- calculate_burnout_rt_tool: Success
- check_utilization_threshold_tool: Success

Dashboard reflects available data only. Full refresh recommended.

---

Integration with Other Skills

Related Skill	Integration Point
production-incident-responder	Escalate when status is RED/BLACK
RESILIENCE_SCORING	Deep-dive into specific metrics
MCP_ORCHESTRATION	Tool failure recovery
safe-schedule-generation	Trigger backup before risky operations
acgme-compliance	Cross-reference compliance violations

---

Configuration

Default parameters (can be overridden via args):

unified_critical_index:
  include_details: true
  top_n: 5

burnout_rt:
  time_window_days: 28

contingency_analysis:
  analyze_n1: true
  analyze_n2: true
  include_cascade_simulation: false
  critical_faculty_only: true

refresh_intervals:
  green: "7 days"
  yellow: "3 days"
  orange: "24 hours"
  red: "4 hours"
  black: "1 hour"

---

Validation Checklist

Before completing resilience dashboard generation, verify:

• All Core Tools Responded: Unified critical index, burnout Rt, utilization, and defense level all returned data
• Status Determination: Overall status (GREEN/YELLOW/ORANGE/RED/BLACK) accurately reflects worst metric severity
• Top Priority Faculty Identified: Critical index includes top N faculty with specific risk patterns
• Actionable Recommendations: Immediate/short-term/long-term actions are specific (not generic)
• Escalation Rules Applied: RED/BLACK status triggers appropriate human escalation warnings
• Partial Failure Handling: If MCP tools failed, dashboard clearly indicates missing data and limitations
• Trend Context Included: If historical data available, trends are shown (improving/worsening)
• Next Refresh Interval: Recommended refresh time matches status severity (GREEN=7d, RED=4h)
• Quality Gate: Dashboard provides enough context for a program coordinator to make decisions

Version

• Created: 2025-12-28
• MCP Tools Required: 8 resilience tools
• Estimated Execution Time: 5-15 seconds (parallel execution)

---

This skill provides a single-pane-of-glass view into program resilience, enabling proactive intervention before issues become crises.