token-waste-eliminationlisted

# Token Waste Elimination Memory files are LLM context, not human documentation. Every line costs tokens. Optimize for my consumption, not human readability. ## Loading Tiers | Tier | What Loads | When | Cost | |------|-----------|------|------| | **Always-on** | `copilot-instructions.md` body, `AGENTS.md`/`CLAUDE.md` (if present), ALL instruction descriptions, ALL skill name+descriptions, ALL agent descriptions | Every request | Fixed | | **Match-loaded** | Instruction bodies with `applyTo` | File context matches glob | Predictable | | **Agent-loaded** | Instruction bodies without `applyTo` | I judge description relevant (not automatic) | Variable | | **On-demand** | Skill bodies, prompt bodies | I load skill when needed; prompt on `/command` | Near-zero | ## Core Rules **Instructions thin, skills thick**: Instructions = decision tables + routing. Skills = procedures + examples + references. **applyTo gates cost**: Add `applyTo` to domain-specific instructions (file-type triggered). Omit for conversational/universal instructions. **No-applyTo != auto-load**: Without `applyTo`, the description is always visible but I choose whether to load the body. Not deterministic. ## Size Thresholds | File Type | Has Matching Skill | Max | Over Threshold | |-----------|-------------------|-----|----------------| | Instruction | Yes | 50 lines | Trim to rules + routing pointer | | Instruction | No | 200 lines | Review for splitting | | Skill body | Any | 400 lines | Move detail t