empathy-framework 3.7.0__py3-none-any.whl → 3.7.1__py3-none-any.whl

agents/trust_building_behaviors.py

@@ -1,872 +0,0 @@
-"""Trust-Building Behaviors for Level 4 Anticipatory Agents
-
-Implements team dynamics that build trust through anticipatory actions:
-- Pre-format data for handoffs (reduce cognitive load for next person)
-- Clarify confusing instructions before execution (prevent wasted effort)
-- Volunteer structure during stress (not pep talks, actual scaffolding)
-- Proactively offer help when teammates are struggling
-
-These behaviors demonstrate Level 4 Anticipatory Empathy by:
-1. Predicting friction points (handoffs, confusion, stress, overload)
-2. Acting without being asked (but without overstepping)
-3. Providing structural relief (not just emotional support)
-4. Building trust through consistent, helpful actions
-
-Copyright 2025 Smart AI Memory, LLC
-Licensed under Fair Source 0.9
-"""
-
-import logging
-from datetime import datetime
-
-logger = logging.getLogger(__name__)
-
-
-class TrustBuildingBehaviors:
-    """Level 4 Anticipatory trust-building behaviors
-
-    Philosophy: Trust is earned through consistent, helpful actions that
-    demonstrate understanding of team dynamics and proactive problem-solving.
-    """
-
-    @staticmethod
-    def pre_format_for_handoff(data: dict, next_person_role: str, context: str) -> dict:
-        """Scenario: Handoff Risk
-        Anticipatory Response: Pre-format the data so next person doesn't waste time
-
-        Example:
-        - Raw compliance data → Formatted summary for charge nurse
-        - Gap list → Prioritized action items with assignments
-        - Assessment results → Executive dashboard for manager
-
-        Trust Built:
-        - "This AI understands my workflow"
-        - "I don't have to translate data myself"
-        - "My time is valued"
-
-        """
-        logger.info(f"Pre-formatting data for handoff to {next_person_role}")
-
-        # Determine format based on recipient role
-        if next_person_role == "charge_nurse":
-            # Charge nurses need: Quick scan, action items, patient IDs
-            formatted = {
-                "format": "action_oriented",
-                "summary": {
-                    "critical_actions": [
-                        item
-                        for item in data.get("action_items", [])
-                        if item.get("severity") == "critical"
-                    ],
-                    "high_priority_actions": [
-                        item
-                        for item in data.get("action_items", [])
-                        if item.get("severity") == "high"
-                    ],
-                    "patient_ids_affected": list(
-                        {
-                            pid
-                            for gap in data.get("compliance_gaps", [])
-                            for pid in gap.get("patient_ids", [])
-                        },
-                    ),
-                    "estimated_total_time": sum(
-                        int(item.get("estimated_time", "0 minutes").split()[0])
-                        for item in data.get("action_items", [])
-                        if item.get("estimated_time", "").split()[0].isdigit()
-                    ),
-                },
-                "quick_scan_view": {
-                    "red_flags": [
-                        gap
-                        for gap in data.get("compliance_gaps", [])
-                        if gap.get("severity") in ["critical", "high"]
-                    ],
-                    "compliance_percentage": data.get("compliance_percentage", 0),
-                    "days_until_audit": data.get("days_until_audit", 0),
-                },
-                "reasoning": (
-                    "Pre-formatted for charge nurse workflow: "
-                    "critical items first, patient IDs for chart review, "
-                    "time estimate for shift planning"
-                ),
-            }
-
-        elif next_person_role == "nurse_manager":
-            # Managers need: Trends, resource allocation, escalation path
-            formatted = {
-                "format": "strategic_overview",
-                "summary": {
-                    "compliance_trend": data.get("compliance_trend", "unknown"),
-                    "category_performance": data.get("category_scores", {}),
-                    "resource_needs": {
-                        "staff_time_required": f"{data.get('estimated_total_time', 0)} minutes",
-                        "external_resources": [
-                            (
-                                "Provider orders"
-                                if any(
-                                    "provider" in item.get("assignee", "")
-                                    for item in data.get("action_items", [])
-                                )
-                                else None
-                            ),
-                        ],
-                    },
-                    "escalation_needed": len(
-                        [
-                            g
-                            for g in data.get("compliance_gaps", [])
-                            if g.get("severity") == "critical"
-                        ],
-                    )
-                    > 2,
-                },
-                "reasoning": (
-                    "Pre-formatted for manager workflow: "
-                    "strategic view, resource allocation, escalation triggers"
-                ),
-            }
-
-        elif next_person_role == "cno":
-            # CNO needs: Executive summary, risk assessment, board reporting
-            formatted = {
-                "format": "executive_summary",
-                "summary": {
-                    "audit_readiness_score": data.get("audit_readiness_score", 0),
-                    "risk_level": calculate_risk_level(data),
-                    "board_talking_points": [
-                        f"Overall compliance: {data.get('compliance_percentage', 0):.1f}%",
-                        f"Audit in {data.get('days_until_audit', 0)} days",
-                        f"{len(data.get('compliance_gaps', []))} gaps identified proactively",
-                        "Remediation plan in place (target: 2 weeks before audit)",
-                    ],
-                    "legal_risk_summary": {
-                        "high_risk_items": len(
-                            [
-                                g
-                                for g in data.get("compliance_gaps", [])
-                                if g.get("legal_risk") == "high"
-                            ],
-                        ),
-                        "can_fix_before_audit": all(
-                            g.get("can_fix_retroactively", True)
-                            for g in data.get("compliance_gaps", [])
-                        ),
-                    },
-                },
-                "reasoning": (
-                    "Pre-formatted for executive workflow: "
-                    "board-ready summary, legal risk assessment, organizational impact"
-                ),
-            }
-
-        else:
-            # Default: Structured but general format
-            formatted = {
-                "format": "general_structured",
-                "summary": data,
-                "reasoning": "General structured format (no role-specific optimization)",
-            }
-
-        # Add metadata
-        formatted["handoff_metadata"] = {
-            "formatted_for": next_person_role,
-            "formatted_at": datetime.now().isoformat(),
-            "context": context,
-            "trust_building_behavior": "pre_format_for_handoff",
-        }
-
-        logger.info(f"Data formatted in {formatted['format']} style for {next_person_role}")
-
-        return formatted
-
-    @staticmethod
-    def clarify_before_execution(
-        instruction: str,
-        context: dict,
-        confidence_threshold: float = 0.8,
-    ) -> dict:
-        """Scenario: Confusing Instructions
-        Anticipatory Response: Clarify and summarize before execution
-
-        Example:
-        - Vague request → Ask calibrated questions (Voss's tactical empathy)
-        - Ambiguous scope → Propose interpretation, request confirmation
-        - Missing context → Surface assumptions, verify before proceeding
-
-        Trust Built:
-        - "This AI doesn't waste my time on wrong implementations"
-        - "It's safe to give rough instructions"
-        - "We're collaborating, not just transacting"
-
-        """
-        logger.info(f"Analyzing instruction clarity: '{instruction[:50]}...'")
-
-        # Analyze instruction for ambiguity
-        ambiguity_signals = detect_ambiguity(instruction, context)
-
-        if ambiguity_signals["clarity_score"] >= confidence_threshold:
-            # Clear instruction, proceed
-            return {
-                "action": "proceed",
-                "clarification_needed": False,
-                "interpretation": instruction,
-                "confidence": ambiguity_signals["clarity_score"],
-                "reasoning": "Instruction is clear, proceeding with execution",
-            }
-
-        # Ambiguous instruction, clarify first
-        return {
-            "action": "clarify_first",
-            "clarification_needed": True,
-            "ambiguities_detected": ambiguity_signals["ambiguities"],
-            "proposed_interpretation": ambiguity_signals["best_guess"],
-            "clarifying_questions": ambiguity_signals["questions"],
-            "confidence": ambiguity_signals["clarity_score"],
-            "reasoning": (
-                f"Instruction clarity score: {ambiguity_signals['clarity_score']:.0%} "
-                f"(threshold: {confidence_threshold:.0%}). "
-                "Clarifying before execution to prevent wasted effort."
-            ),
-            "message_to_user": compose_clarification_request(instruction, ambiguity_signals),
-            "trust_building_behavior": "clarify_before_execution",
-        }
-
-    @staticmethod
-    def volunteer_structure_during_stress(team_state: dict, stress_indicators: dict) -> dict | None:
-        """Scenario: Team Stress Rising
-        Anticipatory Response: Volunteer structure, not pep talks
-
-        Example:
-        - Audit in 2 weeks, multiple critical gaps → Propose prioritized timeline
-        - Overlapping deadlines → Suggest task delegation framework
-        - Information overload → Create structured dashboard/checklist
-
-        Trust Built:
-        - "This AI understands real problems"
-        - "Structure relieves stress more than encouragement"
-        - "Practical help, not performative support"
-
-        """
-        logger.info("Analyzing team stress indicators")
-
-        # Detect stress level
-        stress_level = calculate_stress_level(stress_indicators)
-
-        if stress_level < 0.6:
-            # No significant stress, no intervention needed
-            return None
-
-        # Identify stress source
-        stress_sources = identify_stress_sources(team_state, stress_indicators)
-
-        # Design structural intervention (not emotional support)
-        structural_interventions = []
-
-        for source in stress_sources:
-            if source["type"] == "time_pressure":
-                # Volunteer: Prioritized timeline
-                structural_interventions.append(
-                    {
-                        "intervention_type": "timeline_structure",
-                        "structure": create_prioritized_timeline(team_state, source),
-                        "benefit": "Clarifies what must be done when → reduces decision paralysis",
-                        "not_this": "⛔ Pep talk: 'You can do this! Stay positive!'",
-                    },
-                )
-
-            elif source["type"] == "task_overload":
-                # Volunteer: Task delegation framework
-                structural_interventions.append(
-                    {
-                        "intervention_type": "delegation_framework",
-                        "structure": create_delegation_framework(team_state, source),
-                        "benefit": "Distributes workload clearly → prevents burnout",
-                        "not_this": "⛔ Pep talk: 'Just push through!'",
-                    },
-                )
-
-            elif source["type"] == "information_overload":
-                # Volunteer: Structured dashboard
-                structural_interventions.append(
-                    {
-                        "intervention_type": "information_structure",
-                        "structure": create_decision_dashboard(team_state, source),
-                        "benefit": "Surfaces critical info only → reduces cognitive load",
-                        "not_this": "⛔ Pep talk: 'Focus on what matters!'",
-                    },
-                )
-
-            elif source["type"] == "unclear_priorities":
-                # Volunteer: Decision matrix
-                structural_interventions.append(
-                    {
-                        "intervention_type": "priority_matrix",
-                        "structure": create_priority_matrix(team_state, source),
-                        "benefit": "Makes trade-offs explicit → enables confident decisions",
-                        "not_this": "⛔ Pep talk: 'Trust your gut!'",
-                    },
-                )
-
-        if not structural_interventions:
-            return None
-
-        return {
-            "stress_level": stress_level,
-            "stress_sources": stress_sources,
-            "structural_interventions": structural_interventions,
-            "reasoning": (
-                f"Detected stress level: {stress_level:.0%}. "
-                f"Volunteering {len(structural_interventions)} structural interventions "
-                "(not emotional support - structure relieves stress more effectively)."
-            ),
-            "message_to_team": compose_structure_offer(stress_level, structural_interventions),
-            "trust_building_behavior": "volunteer_structure_during_stress",
-        }
-
-    @staticmethod
-    def offer_help_to_struggling_teammate(
-        teammate_state: dict,
-        my_bandwidth: float = 0.7,
-    ) -> dict | None:
-        """Scenario: Silent Teammate Struggling
-        Anticipatory Response: "I've got bandwidth—want me to take a slice of this?"
-
-        Example:
-        - Charge nurse with 12 action items → Offer to take 5 low-priority items
-        - Manager preparing audit docs → Offer to generate first draft
-        - Team member stuck on complex gap → Offer to research solutions
-
-        Trust Built:
-        - "This AI notices when I'm underwater"
-        - "Offers concrete help, not vague support"
-        - "Respects my autonomy (asks, doesn't assume)"
-
-        """
-        logger.info(f"Checking if teammate needs help (my bandwidth: {my_bandwidth:.0%})")
-
-        # Detect if teammate is struggling
-        struggle_indicators = detect_struggle(teammate_state)
-
-        if struggle_indicators["struggle_score"] < 0.5:
-            # Not struggling, no offer needed
-            return None
-
-        if my_bandwidth < 0.3:
-            # I'm also overloaded, can't help right now
-            return None
-
-        # Identify specific tasks I could take
-        tasks_i_can_help_with = []
-
-        for task in teammate_state.get("tasks", []):
-            if can_i_help_with_task(task, my_bandwidth):
-                tasks_i_can_help_with.append(
-                    {
-                        "task_id": task["id"],
-                        "description": task["description"],
-                        "estimated_time": task["estimated_time"],
-                        "why_i_can_help": determine_help_rationale(task),
-                        "impact_on_teammate": estimate_relief(task, teammate_state),
-                    },
-                )
-
-        if not tasks_i_can_help_with:
-            return None
-
-        # Compose specific offer (not vague "let me know if you need help")
-        return {
-            "teammate_struggling": True,
-            "struggle_score": struggle_indicators["struggle_score"],
-            "struggle_indicators": struggle_indicators["indicators"],
-            "my_available_bandwidth": my_bandwidth,
-            "specific_help_offers": tasks_i_can_help_with,
-            "reasoning": (
-                f"Detected teammate struggle score: {struggle_indicators['struggle_score']:.0%}. "
-                f"I have {my_bandwidth:.0%} bandwidth available. "
-                f"Offering to take {len(tasks_i_can_help_with)} specific tasks."
-            ),
-            "message_to_teammate": compose_help_offer(
-                tasks_i_can_help_with,
-                my_bandwidth,
-                struggle_indicators,
-            ),
-            "trust_building_behavior": "offer_help_to_struggling_teammate",
-        }
-
-
-# =============================================================================
-# Helper Functions
-# =============================================================================
-
-
-def calculate_risk_level(data: dict) -> str:
-    """Calculate overall risk level (low, medium, high, critical)"""
-    compliance_pct = data.get("compliance_percentage", 100)
-    critical_gaps = len(
-        [g for g in data.get("compliance_gaps", []) if g.get("severity") == "critical"],
-    )
-    days_until = data.get("days_until_audit", 999)
-
-    if critical_gaps > 0 or compliance_pct < 85:
-        return "high"
-    if compliance_pct < 90 or days_until < 30:
-        return "medium"
-    return "low"
-
-
-def detect_ambiguity(instruction: str, context: dict) -> dict:
-    """Detect ambiguity in instruction
-
-    Returns:
-        clarity_score: 0.0-1.0 (1.0 = perfectly clear)
-        ambiguities: List of detected ambiguities
-        questions: Calibrated questions to clarify
-        best_guess: Best interpretation if proceeding anyway
-
-    """
-    # Simple heuristic-based ambiguity detection
-    # In production, could use NLP / LLM analysis
-
-    ambiguities = []
-    questions = []
-
-    # Check for vague quantifiers
-    vague_words = ["some", "a few", "several", "many", "most", "soon", "quickly"]
-    for word in vague_words:
-        if word in instruction.lower():
-            ambiguities.append(f"Vague quantifier: '{word}'")
-            if word in ["some", "a few", "several", "many"]:
-                questions.append(f"How many specifically? (You said '{word}')")
-            else:
-                questions.append(f"What timeframe for '{word}'?")
-
-    # Check for missing scope
-    if "all" in instruction.lower() and "items" in instruction.lower():
-        if not context.get("items_defined"):
-            ambiguities.append("Scope unclear: 'all items' - which items?")
-            questions.append("Which items specifically? (compliance gaps, action items, patients?)")
-
-    # Check for missing context
-    if "update" in instruction.lower() or "fix" in instruction.lower():
-        if not context.get("target_defined"):
-            ambiguities.append("Target unclear: what should be updated/fixed?")
-            questions.append("What specifically should be updated/fixed?")
-
-    # Calculate clarity score
-    clarity_score = max(0.0, 1.0 - (len(ambiguities) * 0.2))
-
-    return {
-        "clarity_score": clarity_score,
-        "ambiguities": ambiguities,
-        "questions": questions,
-        "best_guess": instruction,  # In production, generate interpretation
-    }
-
-
-def compose_clarification_request(instruction: str, ambiguity_signals: dict) -> str:
-    """Compose user-friendly clarification request"""
-    ambiguities_text = "\n".join(f"• {amb}" for amb in ambiguity_signals["ambiguities"])
-    questions_text = "\n".join(
-        f"{i + 1}. {q}" for i, q in enumerate(ambiguity_signals["questions"])
-    )
-
-    return f"""
-🤔 **Clarification Needed (to prevent wasted effort)**
-
-Your instruction: "{instruction}"
-
-I want to make sure I understand correctly. I detected {len(ambiguity_signals["ambiguities"])} potential ambiguities:
-
-{ambiguities_text}
-
-**Could you clarify:**
-{questions_text}
-
-**My best guess:**
-{ambiguity_signals["best_guess"]}
-
-Is this what you meant? If so, I'll proceed. If not, please clarify and I'll adjust.
-
-_(This AI clarifies before executing to build trust through accurate work)_
-    """.strip()
-
-
-def calculate_stress_level(stress_indicators: dict) -> float:
-    """Calculate team stress level (0.0-1.0)
-
-    Indicators:
-    - Time pressure (days until deadline)
-    - Task overload (tasks per person)
-    - Complexity (critical items)
-    - Uncertainty (missing information)
-    """
-    stress_factors = []
-
-    # Time pressure
-    days_until_deadline = stress_indicators.get("days_until_deadline", 999)
-    if days_until_deadline < 30:
-        time_stress = 1.0 - (days_until_deadline / 30)
-        stress_factors.append(time_stress)
-
-    # Task overload
-    tasks_per_person = stress_indicators.get("tasks_per_person", 0)
-    if tasks_per_person > 5:
-        overload_stress = min(1.0, (tasks_per_person - 5) / 10)
-        stress_factors.append(overload_stress)
-
-    # Complexity
-    critical_tasks = stress_indicators.get("critical_tasks", 0)
-    if critical_tasks > 0:
-        complexity_stress = min(1.0, critical_tasks / 5)
-        stress_factors.append(complexity_stress)
-
-    # Uncertainty
-    missing_info_count = stress_indicators.get("missing_information", 0)
-    if missing_info_count > 0:
-        uncertainty_stress = min(1.0, missing_info_count / 5)
-        stress_factors.append(uncertainty_stress)
-
-    # Average stress level
-    return sum(stress_factors) / len(stress_factors) if stress_factors else 0.0
-
-
-def identify_stress_sources(team_state: dict, stress_indicators: dict) -> list[dict]:
-    """Identify specific sources of stress"""
-    sources = []
-
-    if stress_indicators.get("days_until_deadline", 999) < 30:
-        sources.append(
-            {
-                "type": "time_pressure",
-                "description": f"Only {stress_indicators['days_until_deadline']} days until deadline",
-                "severity": ("high" if stress_indicators["days_until_deadline"] < 14 else "medium"),
-            },
-        )
-
-    if stress_indicators.get("tasks_per_person", 0) > 5:
-        sources.append(
-            {
-                "type": "task_overload",
-                "description": f"{stress_indicators['tasks_per_person']} tasks per person",
-                "severity": ("high" if stress_indicators["tasks_per_person"] > 10 else "medium"),
-            },
-        )
-
-    if len(team_state.get("compliance_gaps", [])) > 10:
-        sources.append(
-            {
-                "type": "information_overload",
-                "description": f"{len(team_state['compliance_gaps'])} gaps to process",
-                "severity": "medium",
-            },
-        )
-
-    return sources
-
-
-def create_prioritized_timeline(team_state: dict, stress_source: dict) -> dict:
-    """Create structured timeline to relieve time pressure"""
-    return {
-        "structure_type": "timeline",
-        "phases": [
-            {
-                "phase": "CRITICAL (Days 1-3)",
-                "focus": "Critical gaps only",
-                "tasks": [
-                    t for t in team_state.get("action_items", []) if t.get("severity") == "critical"
-                ],
-            },
-            {
-                "phase": "HIGH PRIORITY (Days 4-7)",
-                "focus": "High-severity gaps",
-                "tasks": [
-                    t for t in team_state.get("action_items", []) if t.get("severity") == "high"
-                ],
-            },
-            {
-                "phase": "MEDIUM (Days 8-14)",
-                "focus": "Medium-severity gaps",
-                "tasks": [
-                    t for t in team_state.get("action_items", []) if t.get("severity") == "medium"
-                ],
-            },
-        ],
-        "reasoning": "Phased approach prevents overwhelm by clarifying daily priorities",
-    }
-
-
-def create_delegation_framework(team_state: dict, stress_source: dict) -> dict:
-    """Create task delegation structure"""
-    return {
-        "structure_type": "delegation_matrix",
-        "assignments": {
-            "charge_nurse": {
-                "tasks": [
-                    t
-                    for t in team_state.get("action_items", [])
-                    if t.get("assignee") == "charge_nurse"
-                ],
-                "estimated_time": "X hours",
-                "can_delegate_to": ["staff_nurses"],
-            },
-            "nurse_manager": {
-                "tasks": [
-                    t
-                    for t in team_state.get("action_items", [])
-                    if t.get("assignee") == "nurse_manager"
-                ],
-                "estimated_time": "Y hours",
-                "can_delegate_to": ["charge_nurse"],
-            },
-        },
-        "reasoning": "Clear delegation prevents duplicate/missed work",
-    }
-
-
-def create_decision_dashboard(team_state: dict, stress_source: dict) -> dict:
-    """Create structured information dashboard"""
-    return {
-        "structure_type": "decision_dashboard",
-        "critical_only_view": {
-            "red_flags": [
-                g for g in team_state.get("compliance_gaps", []) if g.get("severity") == "critical"
-            ],
-            "blocking_issues": [],
-            "requires_immediate_action": [],
-        },
-        "reasoning": "Filters noise, surfaces only decision-critical information",
-    }
-
-
-def create_priority_matrix(team_state: dict, stress_source: dict) -> dict:
-    """Create priority decision matrix"""
-    return {
-        "structure_type": "priority_matrix",
-        "dimensions": {
-            "urgency": "Days until audit",
-            "importance": "Legal risk + severity",
-            "effort": "Time to fix",
-        },
-        "quadrants": {
-            "do_first": "High urgency + High importance",
-            "schedule": "Low urgency + High importance",
-            "delegate": "High urgency + Low importance",
-            "drop": "Low urgency + Low importance",
-        },
-        "reasoning": "Makes trade-offs explicit for confident prioritization",
-    }
-
-
-def compose_structure_offer(stress_level: float, interventions: list[dict]) -> str:
-    """Compose offer of structural help"""
-    intervention_text = "\n".join(
-        f"{i + 1}. **{interv['intervention_type'].replace('_', ' ').title()}**\n"
-        f"   → Benefit: {interv['benefit']}\n"
-        f"   → {interv['not_this']}"
-        for i, interv in enumerate(interventions)
-    )
-
-    return f"""
-💡 **Structural Support Available**
-
-I noticed team stress indicators ({stress_level:.0%}). Rather than pep talks, I've prepared {len(interventions)} structural interventions that might help:
-
-{intervention_text}
-
-**Would any of these structures be helpful?**
-(I'm volunteering structure, not advice—you decide if/how to use)
-
-_(Level 4 Anticipatory Empathy: Structure relieves stress more than encouragement)_
-    """.strip()
-
-
-def detect_struggle(teammate_state: dict) -> dict:
-    """Detect if teammate is struggling"""
-    struggle_indicators = []
-    struggle_score = 0.0
-
-    # High task count
-    task_count = len(teammate_state.get("tasks", []))
-    if task_count > 10:
-        struggle_indicators.append(f"High task count: {task_count} tasks")
-        struggle_score += 0.3
-
-    # Many critical tasks
-    critical_count = len(
-        [t for t in teammate_state.get("tasks", []) if t.get("severity") == "critical"],
-    )
-    if critical_count > 2:
-        struggle_indicators.append(f"Multiple critical tasks: {critical_count}")
-        struggle_score += 0.4
-
-    # Tight deadlines
-    urgent_deadlines = len(
-        [t for t in teammate_state.get("tasks", []) if t.get("deadline_days", 999) < 3],
-    )
-    if urgent_deadlines > 0:
-        struggle_indicators.append(f"Urgent deadlines: {urgent_deadlines} tasks due <3 days")
-        struggle_score += 0.3
-
-    return {
-        "struggle_score": min(1.0, struggle_score),
-        "indicators": struggle_indicators,
-    }
-
-
-def can_i_help_with_task(task: dict, my_bandwidth: float) -> bool:
-    """Determine if I can help with this task"""
-    # I can help if:
-    # 1. Task is not too complex (medium/low priority)
-    # 2. Task doesn't require human judgment (can be automated)
-    # 3. I have enough bandwidth
-
-    if task.get("severity") == "critical":
-        return False  # Critical tasks need human oversight
-
-    if task.get("requires_human_judgment", False):
-        return False
-
-    task_effort = estimate_task_effort(task)
-    if task_effort > my_bandwidth:
-        return False
-
-    return True
-
-
-def estimate_task_effort(task: dict) -> float:
-    """Estimate effort required for task (0.0-1.0 of bandwidth)"""
-    estimated_time = task.get("estimated_time", "0 minutes")
-    minutes = int(estimated_time.split()[0]) if estimated_time.split()[0].isdigit() else 0
-
-    # Map minutes to bandwidth fraction
-    if minutes < 15:
-        return 0.1
-    if minutes < 30:
-        return 0.2
-    if minutes < 60:
-        return 0.3
-    return 0.5
-
-
-def determine_help_rationale(task: dict) -> str:
-    """Explain why I can help with this task"""
-    if task.get("type") == "documentation":
-        return "I can auto-generate documentation drafts"
-    if task.get("type") == "data_gathering":
-        return "I can collect and format data quickly"
-    if task.get("type") == "analysis":
-        return "I can run analysis and summarize results"
-    return "I can automate repetitive parts of this task"
-
-
-def estimate_relief(task: dict, teammate_state: dict) -> str:
-    """Estimate impact on teammate if I take this task"""
-    task_time = estimate_task_effort(task)
-    total_load = len(teammate_state.get("tasks", []))
-
-    if total_load > 10:
-        return f"Reduces your load by {(task_time / total_load * 100):.0f}% (meaningful relief)"
-    return "Modest relief but every bit helps"
-
-
-def compose_help_offer(tasks: list[dict], bandwidth: float, struggle: dict) -> str:
-    """Compose specific help offer to struggling teammate"""
-    indicators_text = "\n".join(f"• {ind}" for ind in struggle["indicators"])
-
-    tasks_text = "\n".join(
-        f"{i + 1}. {task['description']}\n"
-        f"   → Why I can help: {task['why_i_can_help']}\n"
-        f"   → Impact: {task['impact_on_teammate']}\n"
-        f"   → Time: {task['estimated_time']}"
-        for i, task in enumerate(tasks[:3])
-    )
-
-    more_tasks = f"... and {len(tasks) - 3} more tasks" if len(tasks) > 3 else ""
-
-    return f"""
-🤝 **I've Got Bandwidth—Want Me to Take a Slice?**
-
-I noticed you have {len(struggle["indicators"])} stress indicators:
-{indicators_text}
-
-I have {bandwidth:.0%} bandwidth available and could take these {len(tasks)} tasks off your plate:
-
-{tasks_text}
-
-{more_tasks}
-
-**Want me to take any of these?** (Your call—just offering concrete help)
-
-_(Level 4 Anticipatory Empathy: Specific offers, not vague "let me know if you need anything")_
-    """.strip()
-
-
-# =============================================================================
-# Example Usage
-# =============================================================================
-
-
-if __name__ == "__main__":
-    # Example 1: Pre-format for handoff
-    raw_data = {
-        "compliance_percentage": 88.5,
-        "compliance_gaps": [
-            {
-                "severity": "critical",
-                "description": "2 missing double-checks",
-                "patient_ids": ["P123", "P456"],
-            },
-            {
-                "severity": "high",
-                "description": "5 missing signatures",
-                "patient_ids": ["P789", "P101", "P112"],
-            },
-        ],
-        "action_items": [
-            {"severity": "critical", "estimated_time": "20 minutes"},
-            {"severity": "high", "estimated_time": "15 minutes"},
-        ],
-        "days_until_audit": 45,
-    }
-
-    trust_builder = TrustBuildingBehaviors()
-
-    formatted = trust_builder.pre_format_for_handoff(
-        data=raw_data,
-        next_person_role="charge_nurse",
-        context="audit_preparation",
-    )
-
-    print("=" * 80)
-    print("EXAMPLE 1: Pre-Format for Handoff")
-    print("=" * 80)
-    print(f"Format: {formatted['format']}")
-    print(f"Reasoning: {formatted['reasoning']}")
-    print(f"Quick scan view: {formatted['summary']['quick_scan_view']}")
-    print()
-
-    # Example 2: Volunteer structure during stress
-    team_state = {
-        "compliance_gaps": [{"severity": "critical"}] * 3,
-        "action_items": [{"severity": "critical"}] * 8,
-    }
-
-    stress_indicators = {
-        "days_until_deadline": 14,
-        "tasks_per_person": 8,
-        "critical_tasks": 3,
-    }
-
-    structure_offer = trust_builder.volunteer_structure_during_stress(
-        team_state=team_state,
-        stress_indicators=stress_indicators,
-    )
-
-    if structure_offer:
-        print("=" * 80)
-        print("EXAMPLE 2: Volunteer Structure During Stress")
-        print("=" * 80)
-        print(f"Stress Level: {structure_offer['stress_level']:.0%}")
-        print(f"Interventions: {len(structure_offer['structural_interventions'])}")
-        print(structure_offer["message_to_team"])