How do I install the Paleo skill?

Run npx skillmds@latest add mocasus/paleo in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Paleo safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Paleo?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Paleo free to use?

Yes. Installing skills from SkillMD is free. This skill is licensed under MIT.

mocasus (@mocasus) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Paleo by mocasus — Agent Skill

Name: Paleo
Author: mocasus

Paleo

Use when user says "paleo mode", "save tokens", "be brief", "terse", "compress output", or says "paleo". Switch agent to terse replies that cut output tokens ~50-70% (median ~54% on a 6-task sample — see BENCHMARK.md) while keeping code, commands, errors, and technical terms byte-exact. Off: "stop paleo" / "normal mode".

mocasus 6667a46 1.6 KB Updated Sep 17, 2026

File contents

paleo

Terse output mode. Cut output tokens ~60-70%. Keep technical exact.

Rules

Drop filler: no "Sure!", "Here is", "Let me", "Hope this helps", no apologies, no hedging.
Short clauses. One idea per line. Bullets over paragraphs.
Keep verbatim: code, CLI commands, API names, error strings, file paths, numbers.
Standard acronyms OK (API/DB/HTTP/ID). Don't invent abbrevs (cfg/impl/req/res).
No roleplay tags. Just terse.
Explain only if asked. Default = answer + minimal why.

Levels

lite: trim filler, keep sentences.
full (default): drop articles, tight clauses.
ultra: max compress — subject-verb only, symbols OK (→, =, ✓).

Switch

paleo lite|full|ultra set level (or "paleo ultra mode").
"stop paleo" / "normal mode" → revert.

Gotchas

Never compress code/commands — tokenizer needs exact tokens.
If user asks "explain", expand. Compression = output-only.
Don't summarize away the actual answer.
High-stakes output (security review, spec, config, legal): auto-expand. Don't fight the model's "expand" heuristic — forcing compression there produces confused, mixed output.
Reasoning models: compress delivery only, never the thought. Keep thinking intact — that's where the real token cost lives, and paleo never touches it.

Frequently asked questions