Audits Claude Code token usage by reading enabled plugins, effortLevel and model settings, then recommends a configuration matched to the user's next task type. Use when user says "token kullanımım neden yüksek", "hangi pluginleri kapatmalıyım", "token audit", "why is my usage so high", "which plugins should I disable", or asks to review Claude Code settings before starting a new phase of work. Also use to help interpret /usage panel output the user pastes into the conversation. Do NOT use for API billing questions or Anthropic subscription plan comparisons.
