moai-learning-optimizer

What It Does

Intelligent learning system optimizer that analyzes Alfred's session patterns, identifies improvement opportunities, and continuously enhances performance through adaptive learning and knowledge optimization. Learns from user interactions to provide increasingly relevant and efficient assistance. Core capabilities include session pattern analysis, performance optimization based on usage, adaptive skill selection and invocation timing, knowledge gap identification and filling, user preference learning and personalization, system performance monitoring and tuning, predictive assistance, and continuous improvement through feedback integration.

How it works

The skill provides concrete components and Python code blocks that perform:

• Session Pattern Analysis: collects metrics like session_duration, tool_usage_patterns, skill_invocation_patterns, user_interactions, success_rates, bottlenecks, and satisfaction; identifies patterns such as peak_times, preferred_tools, skill_effectiveness, and workflow_optimization; returns metrics, patterns, and learning recommendations.
• Adaptive Learning System: initializes knowledge_base, user_preferences, performance_history; learns from a session by extracting signals, updating knowledge, adjusting preferences, optimizing parameters, and saving state; predicts needs given a context by returning likely_tools, optimal_skills, potential_issues, and recommended_actions; adapts responses based on user_feedback by analyzing feedback, adjusting strategies, updating communication preferences, and refining assistance.
• Performance Optimization Engine: computes an optimization plan across areas like skill_loading, response_time, context_usage, knowledge_retrieval, tool_selection, and workflow_efficiency; builds a plan with current and target performance, strategies, priority, and expected improvements.
• Knowledge Management: analyzes knowledge gaps for missing or outdated information, prioritizes gaps, and generates a learning plan with immediate, medium-term, and long-term needs plus resources and strategy. Integrates new knowledge with steps for validation, categorization, indexing, linking, testing, and deployment; maintains knowledge quality through metrics (accuracy, relevance, completeness, consistency, freshness) and executes improvements when issues exist.
• User Personalization: learns user preferences (communication_style, detail_level, tool_preferences, workflows, timing, error_handling); builds a user profile and personalizes Alfred's behavior. Adaptive Assistance adjusts intervention level, content detail, and explanation style based on context; personalizes responses in content, tone, format, and timing.
• Experience Optimization: measures response_satisfaction, task_completion_efficiency, learning_progress, error_recovery_time, and engagement; identifies improvements and creates an optimization plan with current/target state, improvements, timeline, and success metrics.
• Predictive Analytics: Behavior Prediction forecasts likely actions, issues, interventions, and resource needs based on behavioral patterns and context; Performance Prediction estimates duration, bottlenecks, resource needs, and success probability using history and task features; Optimization Opportunities identifies skill and workflow improvements, performance tuning, and knowledge enhancements, then prioritizes them.
• Continuous Improvement: integrates user feedback to update responses, workflows, knowledge, and performance; manages the learning loop with data collection, pattern analysis, insights, improvements, validation, and state updates; system evolution evolves capabilities using a roadmap and resource planning, implementing steps and validating outcomes.

When to Use

Use when optimizing Alfred's performance and behavior, during session analysis and pattern discovery, when implementing adaptive learning capabilities, for system performance monitoring and tuning, when personalizing Alfred's responses and recommendations, during troubleshooting, and for continuous system improvement and optimization.

What It Can Touch

The skill lists allowed tools including Read, Glob, Grep, Bash, Write, and utilizes the claude-code tool for implementation steps across analyses, learning, and optimization routines.

Caveats

License is MIT. Status is stable. The skill defines learning loops and adaptive mechanisms that rely on analyzed session data and feedback; outcomes depend on data quality and context and are not guaranteed as explicit results.