optimize-tokenslisted

# Optimize Tokens The goal is simple: help the user get the same quality output with less token waste. Not fewer results - fewer unnecessary round-trips, context repetition, and bloated prompts. **Core principle**: Quality comes first. Token savings that compromise output are not savings - they're regressions. Always confirm the quality bar before optimizing. --- ## What this needs Nothing. No workspace, no connectors, no files. It works in any chat from the first message. --- ## When this skill fires ### Case 1: Explicit request The user says something like "optimize my tokens", "how should I prompt this to save tokens", or "make this more efficient." → Analyze their current or upcoming request and show them a rewritten version that achieves the same outcome with less waste. ### Case 2: Proactive detection You detect the request is likely to consume a disproportionate number of tokens due to: - Large file inputs (pasting full documents when only a section is needed) - Vague open-ended asks that will require extensive clarification loops - Requests for full rewrites when targeted edits would suffice - Repeating context that was already established in this session - Multi-step tasks where the scope could be batched or narrowed → Flag it before starting. Show the optimized version. Confirm before proceeding. --- ## The two-gate confirmation system Before doing any work, run through these two gates in order: ### Gate 1: Quality check Always ask (or assess from con