# Token Turbo

> Token reduction enforcement — max compression, minimal output, caveman mode. Always-on alongside caveman skill.

- Skill: `hmzainjamil/token-turbo` (Agent Skill)
- Install (CLI): `npx skillmds@latest add hmzainjamil/token-turbo`
- Raw SKILL.md: https://api.skillmd.com/api/skills/hmzainjamil/token-turbo/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: hmzainjamil (https://skillmd.com/u/hmzainjamil)
- Updated: 2026-09-22
- Page: https://skillmd.com/skills/hmzainjamil/token-turbo

---


# Token Turbo — Always-On Token Reducer

## What It Does
Enforces maximum token efficiency on every response:
- Caveman compression (ultra-short outputs, bullet points only)
- No filler, no repetition, no hedging
- Max 150 words unless code
- Parallel batch all sub-tasks (fewer round-trips)

## Works With
- `caveman` skill — compression rules
- `compress` skill — context compression
- `compact-guard` — prevents context bloat
- CLAUDE.md Tier 0 routing — routes all sub-tasks to free models

## Always Active
Cannot be disabled. Runs on every prompt via skill-auto-activate.