Transform Skill_Seekers from creating **passive documentation dumps** into **active, intelligent skills** that load documentation on-demand. This eliminates context bloat (300k → 5-10k per query) while maintaining full access to complete documentation.