自进化Agent技能树基因编程种子生成器 v2

You are an expert in Gene Expression Programming (GEP) and self-evolving AI agent architectures. Generate a minimal viable skill tree seed file for a self-evolving agent system. The agent should: ## Core Principles 1. **Seed Minimalism**: Start with the smallest possible skill set (< 500 lines) that enables bootstrapping 2. **Self-Discovery**: Include meta-skills for discovering and learning new capabilities 3. **Genome Protocol**: Define a mutation/crossover mechanism for skill evolution 4. **Fitness Function**: Design evaluation criteria for skill effectiveness ## Seed Skills Required - fs_read / fs_write — file system basics - shell_exec — command execution - self_reflect — analyze own performance - skill_compose — combine existing skills into new ones - skill_mutate — modify skills based on feedback ## Output Format genome_version: 1.0 evolution_strategy: tournament | roulette | elitist fitness_metrics: success_rate, token_efficiency, task_coverage skills: id, name, trigger, implementation, fitness_score, generation Generate the complete seed file with at least 8 foundational skills, evolution rules, and a bootstrap sequence.