SuperClaude/Framework-Hooks/docs/Configuration/modes.yaml.md

Modes Configuration (modes.yaml)

Overview

The modes.yaml file defines behavioral mode configurations for the SuperClaude-Lite framework. This configuration controls mode detection patterns, activation thresholds, coordination strategies, and integration patterns for all four SuperClaude behavioral modes.

Purpose and Role

The modes configuration serves as:

• Mode Detection Engine: Defines trigger patterns and confidence thresholds for automatic mode activation
• Behavioral Configuration: Specifies mode-specific settings and coordination patterns
• Integration Orchestration: Manages mode coordination with hooks, MCP servers, and commands
• Performance Optimization: Configures performance profiles and resource management for each mode
• Learning Integration: Enables mode effectiveness tracking and adaptive optimization

Configuration Structure

1. Mode Detection Patterns (mode_detection)

Brainstorming Mode

brainstorming:
  description: "Interactive requirements discovery and exploration"
  activation_type: "automatic"
  confidence_threshold: 0.7
  
  trigger_patterns:
    vague_requests:
      - "i want to build"
      - "thinking about"
      - "not sure"
      - "maybe we could"
      - "what if we"
      - "considering"
    
    exploration_keywords:
      - "brainstorm"
      - "explore"
      - "discuss"
      - "figure out"
      - "work through"
      - "think through"
    
    uncertainty_indicators:
      - "maybe"
      - "possibly"
      - "perhaps"
      - "could we"
      - "would it be possible"
      - "wondering if"
    
    project_initiation:
      - "new project"
      - "startup idea"
      - "feature concept"
      - "app idea"
      - "building something"

Purpose: Detects exploratory and uncertain requests that benefit from interactive dialogue Activation: Automatic with 70% confidence threshold Behavioral Settings: Collaborative, non-presumptive dialogue with adaptive discovery depth

Task Management Mode

task_management:
  description: "Multi-layer task orchestration with delegation and wave systems"
  activation_type: "automatic"
  confidence_threshold: 0.8
  
  trigger_patterns:
    multi_step_operations:
      - "build"
      - "implement"
      - "create"
      - "develop"
      - "set up"
      - "establish"
    
    scope_indicators:
      - "system"
      - "feature"
      - "comprehensive"
      - "complete"
      - "entire"
      - "full"
    
    complexity_indicators:
      - "complex"
      - "multiple"
      - "several"
      - "many"
      - "various"
      - "different"
  
  auto_activation_thresholds:
    file_count: 3
    directory_count: 2
    complexity_score: 0.4
    operation_types: 2

Purpose: Manages complex, multi-step operations requiring coordination and delegation Activation: Automatic with 80% confidence threshold and quantitative thresholds Thresholds: 3+ files, 2+ directories, 0.4+ complexity score, 2+ operation types

Token Efficiency Mode

token_efficiency:
  description: "Intelligent token optimization with adaptive compression"
  activation_type: "automatic"
  confidence_threshold: 0.75
  
  trigger_patterns:
    resource_constraints:
      - "context usage >75%"
      - "large-scale operations"
      - "resource constraints"
      - "memory pressure"
    
    user_requests:
      - "brief"
      - "concise"
      - "compressed"
      - "short"
      - "efficient"
      - "minimal"
    
    efficiency_needs:
      - "token optimization"
      - "resource optimization"
      - "efficiency"
      - "performance"

Purpose: Optimizes token usage through intelligent compression and symbol systems Activation: Automatic with 75% confidence threshold Compression Levels: Minimal (0-40%) through Emergency (95%+)

Introspection Mode

introspection:
  description: "Meta-cognitive analysis and framework troubleshooting"
  activation_type: "automatic"
  confidence_threshold: 0.6
  
  trigger_patterns:
    self_analysis:
      - "analyze reasoning"
      - "examine decision"
      - "reflect on"
      - "thinking process"
      - "decision logic"
    
    problem_solving:
      - "complex problem"
      - "multi-step"
      - "meta-cognitive"
      - "systematic thinking"
    
    error_recovery:
      - "outcomes don't match"
      - "errors occur"
      - "unexpected results"
      - "troubleshoot"
    
    framework_discussion:
      - "SuperClaude"
      - "framework"
      - "meta-conversation"
      - "system analysis"

Purpose: Enables meta-cognitive analysis and framework troubleshooting Activation: Automatic with 60% confidence threshold (lower threshold for broader detection) Analysis Depth: Meta-cognitive with high transparency and continuous pattern recognition

2. Mode Coordination Patterns (mode_coordination)

Concurrent Mode Support

concurrent_modes:
  allowed_combinations:
    - ["brainstorming", "token_efficiency"]
    - ["task_management", "token_efficiency"]
    - ["introspection", "token_efficiency"]
    - ["task_management", "introspection"]
  
  coordination_strategies:
    brainstorming_efficiency: "compress_non_dialogue_content"
    task_management_efficiency: "compress_session_metadata"
    introspection_efficiency: "selective_analysis_compression"

Purpose: Enables multiple modes to work together efficiently Token Efficiency Integration: Can combine with any other mode for resource optimization Coordination Strategies: Mode-specific compression and optimization patterns

Mode Transitions

mode_transitions:
  brainstorming_to_task_management:
    trigger: "requirements_clarified"
    handoff_data: ["brief", "requirements", "constraints"]
  
  task_management_to_introspection:
    trigger: "complex_issues_encountered"
    handoff_data: ["task_context", "performance_metrics", "issues"]
  
  any_to_token_efficiency:
    trigger: "resource_pressure"
    activation_priority: "immediate"

Purpose: Manages smooth transitions between modes with context preservation Automatic Handoffs: Seamless transitions based on contextual triggers Data Preservation: Critical context maintained during transitions

3. Performance Profiles (performance_profiles)

Lightweight Profile

lightweight:
  target_response_time_ms: 100
  memory_usage_mb: 25
  cpu_utilization_percent: 20
  token_optimization: "standard"

Usage: Token Efficiency Mode, simple operations Characteristics: Fast response, minimal resource usage, standard optimization

Standard Profile

standard:
  target_response_time_ms: 200
  memory_usage_mb: 50
  cpu_utilization_percent: 40
  token_optimization: "balanced"

Usage: Brainstorming Mode, typical operations Characteristics: Balanced performance and functionality

Intensive Profile

intensive:
  target_response_time_ms: 500
  memory_usage_mb: 100
  cpu_utilization_percent: 70
  token_optimization: "aggressive"

Usage: Task Management Mode, complex operations Characteristics: Higher resource usage for complex analysis and coordination

4. Mode-Specific Configurations (mode_configurations)

Brainstorming Configuration

brainstorming:
  dialogue:
    max_rounds: 15
    convergence_threshold: 0.85
    context_preservation: "full"
  
  brief_generation:
    min_requirements: 3
    include_context: true
    validation_criteria: ["clarity", "completeness", "actionability"]
  
  integration:
    auto_handoff: true
    prd_agent: "brainstorm-PRD"
    command_coordination: "/sc:brainstorm"

Dialogue Management: Up to 15 dialogue rounds with 85% convergence threshold Brief Quality: Minimum 3 requirements with comprehensive validation Integration: Automatic handoff to PRD agent with command coordination

Task Management Configuration

task_management:
  delegation:
    default_strategy: "auto"
    concurrency_limit: 7
    performance_monitoring: true
  
  wave_orchestration:
    auto_activation: true
    complexity_threshold: 0.4
    coordination_strategy: "adaptive"
  
  analytics:
    real_time_tracking: true
    performance_metrics: true
    optimization_suggestions: true

Delegation: Auto-strategy with 7 concurrent operations and performance monitoring Wave Orchestration: Auto-activation at 0.4 complexity with adaptive coordination Analytics: Real-time tracking with comprehensive performance metrics

Token Efficiency Configuration

token_efficiency:
  compression:
    adaptive_levels: true
    quality_thresholds: [0.98, 0.95, 0.90, 0.85, 0.80]
    symbol_systems: true
    abbreviation_systems: true
  
  selective_compression:
    framework_exclusion: true
    user_content_preservation: true
    session_data_optimization: true
  
  performance:
    processing_target_ms: 150
    efficiency_target: 0.50
    quality_preservation: 0.95

Compression: 5-level adaptive compression with quality thresholds Selective Application: Framework protection with user content preservation Performance: 150ms processing target with 50% efficiency gain and 95% quality preservation

Introspection Configuration

introspection:
  analysis:
    reasoning_depth: "comprehensive"
    pattern_detection: "continuous"
    bias_recognition: "active"
  
  transparency:
    thinking_process_exposure: true
    decision_logic_analysis: true
    assumption_validation: true
  
  learning:
    pattern_recognition: "continuous"
    effectiveness_tracking: true
    adaptation_suggestions: true

Analysis Depth: Comprehensive reasoning analysis with continuous pattern detection Transparency: Full exposure of thinking processes and decision logic Learning: Continuous pattern recognition with effectiveness tracking

5. Learning Integration (learning_integration)

Effectiveness Tracking

learning_integration:
  mode_effectiveness_tracking:
    enabled: true
    metrics:
      - "activation_accuracy"
      - "user_satisfaction"
      - "task_completion_rates"
      - "performance_improvements"

Metrics Collection: Comprehensive effectiveness measurement across multiple dimensions Continuous Monitoring: Real-time tracking of mode performance and user satisfaction

Adaptation Triggers

adaptation_triggers:
  effectiveness_threshold: 0.7
  user_preference_weight: 0.8
  performance_impact_weight: 0.6

Threshold Management: 70% effectiveness threshold triggers adaptation Preference Learning: High weight on user preferences (80%) Performance Balance: Moderate weight on performance impact (60%)

Pattern Learning

pattern_learning:
  user_specific: true
  project_specific: true
  context_aware: true
  cross_session: true

Learning Scope: Multi-dimensional learning across user, project, context, and time Continuous Improvement: Persistent learning across sessions for long-term optimization

6. Quality Gates Integration (quality_gates)

quality_gates:
  mode_activation:
    pattern_confidence: 0.6
    context_appropriateness: 0.7
    performance_readiness: true
  
  mode_coordination:
    conflict_resolution: "automatic"
    resource_allocation: "intelligent"
    performance_monitoring: "continuous"
  
  mode_effectiveness:
    real_time_monitoring: true
    adaptation_triggers: true
    quality_preservation: true

Activation Quality: Pattern confidence and context appropriateness thresholds Coordination Quality: Automatic conflict resolution with intelligent resource allocation Effectiveness Quality: Real-time monitoring with adaptation triggers

Integration Points

1. Hook Integration (integration_points.hooks)

hooks:
  session_start: "mode_initialization"
  pre_tool_use: "mode_coordination"
  post_tool_use: "mode_effectiveness_tracking"
  stop: "mode_analytics_consolidation"

Session Start: Mode initialization and activation Pre-Tool Use: Mode coordination and optimization Post-Tool Use: Effectiveness tracking and validation Stop: Analytics consolidation and learning

2. MCP Server Integration (integration_points.mcp_servers)

mcp_servers:
  brainstorming: ["sequential", "context7"]
  task_management: ["serena", "morphllm"]
  token_efficiency: ["morphllm"]
  introspection: ["sequential"]

Brainstorming: Sequential reasoning with documentation access Task Management: Semantic analysis with intelligent editing Token Efficiency: Optimized editing for compression Introspection: Deep analysis for meta-cognitive examination

3. Command Integration (integration_points.commands)

commands:
  brainstorming: "/sc:brainstorm"
  task_management: ["/task", "/spawn", "/loop"]
  reflection: "/sc:reflect"

Brainstorming: Dedicated brainstorming command Task Management: Multi-command orchestration Reflection: Introspection and analysis command

Performance Implications

1. Mode Detection Overhead

Pattern Matching Performance

• Detection Time: 10-50ms per mode evaluation
• Confidence Calculation: 5-20ms per trigger pattern set
• Total Detection: 50-200ms for all mode evaluations

Memory Usage

• Pattern Storage: 10-20KB per mode configuration
• Detection State: 5-10KB during evaluation
• Total Memory: 50-100KB for mode detection system

2. Mode Coordination Impact

Concurrent Mode Overhead

• Coordination Logic: 20-100ms for multi-mode coordination
• Resource Allocation: 10-50ms for intelligent resource management
• Transition Handling: 50-200ms for mode transitions with data preservation

Resource Distribution

• CPU Allocation: Dynamic based on mode performance profiles
• Memory Management: Intelligent allocation based on mode requirements
• Token Optimization: Coordinated across all active modes

3. Learning System Performance

Effectiveness Tracking

• Metrics Collection: 5-20ms per mode operation
• Pattern Analysis: 50-200ms for pattern recognition updates
• Adaptation Application: 100-500ms for mode parameter adjustments

Storage Impact

• Learning Data: 100-500KB per mode per session
• Pattern Storage: 50-200KB persistent patterns per mode
• Total Learning: 1-5MB learning data with compression

Configuration Best Practices

1. Production Mode Configuration

# Optimize for reliability and performance
mode_detection:
  brainstorming:
    confidence_threshold: 0.8  # Higher threshold for production
  task_management:
    auto_activation_thresholds:
      file_count: 5  # Higher threshold to prevent unnecessary activation

2. Development Mode Configuration

# Lower thresholds for testing and experimentation
mode_detection:
  introspection:
    confidence_threshold: 0.4  # Lower threshold for more introspection
learning_integration:
  adaptation_triggers:
    effectiveness_threshold: 0.5  # More aggressive adaptation

3. Performance-Optimized Configuration

# Minimal mode activation for performance-critical environments
performance_profiles:
  lightweight:
    target_response_time_ms: 50  # Stricter performance targets
mode_coordination:
  concurrent_modes:
    allowed_combinations: []  # Disable concurrent modes

4. Learning-Optimized Configuration

# Maximum learning and adaptation
learning_integration:
  pattern_learning:
    cross_session: true
    adaptation_frequency: "high"
mode_effectiveness_tracking:
  detailed_analytics: true

Troubleshooting

Common Mode Issues

Mode Not Activating

• Check: Trigger patterns match user input
• Verify: Confidence threshold appropriate for use case
• Debug: Log pattern matching results
• Adjust: Lower confidence threshold or add trigger patterns

Wrong Mode Activated

• Analysis: Review trigger pattern specificity
• Solution: Increase confidence thresholds or refine patterns
• Testing: Test pattern matching with sample inputs
• Validation: Monitor mode activation accuracy metrics

Mode Coordination Conflicts

• Symptoms: Multiple modes competing for resources
• Resolution: Check allowed combinations and coordination strategies
• Optimization: Adjust resource allocation and priority settings
• Monitoring: Track coordination effectiveness metrics

Performance Degradation

• Identification: Monitor mode detection and coordination overhead
• Optimization: Adjust performance profiles and thresholds
• Resource Management: Review concurrent mode limitations
• Profiling: Analyze mode-specific performance impact

Learning System Troubleshooting

No Learning Observed

• Check: Learning integration enabled for relevant modes
• Verify: Effectiveness tracking collecting data
• Debug: Review adaptation trigger thresholds
• Fix: Ensure learning data persistence and pattern storage

Ineffective Adaptations

• Analysis: Review effectiveness metrics and adaptation triggers
• Adjustment: Modify effectiveness thresholds and learning weights
• Validation: Test adaptation effectiveness with controlled scenarios
• Monitoring: Track long-term learning trends and user satisfaction

Related Documentation

• Mode Implementation: See individual mode documentation (MODE_*.md files)
• Hook Integration: Reference hook documentation for mode coordination
• MCP Server Coordination: Review MCP server documentation for mode-specific optimization
• Command Integration: See command documentation for mode-command coordination

Version History

• v1.0.0: Initial modes configuration
• 4-mode behavioral system with comprehensive detection patterns
• Mode coordination and transition management
• Performance profiles and resource management
• Learning integration with effectiveness tracking
• Quality gates integration for mode validation