Self-Improving Agent "An AI agent that learns from every interaction, accumulating patterns and insights to continuously improve its own capabilities." — Based on 2025 lifelong learning research Overview This is a universal self-improvement system that learns from ALL skill experiences, not just PRDs. It implements a complete feedback loop with: - Multi-Memory Architecture : Semantic + Episodic + Working memory - Self-Correction : Detects and fixes skill guidance errors - Self-Validation : Periodically verifies skill accuracy - Hooks Integration : Auto-triggers on skill events (before start,…