一键在 Manus 中运行任何 Skill

paleo

星标17

分支0

更新时间2026年7月10日 18:09

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".

安装

用 Codex 或 Claude 帮你安装复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。

在 Manus 中运行

来源

mocasus

mocasus/paleo

打开 GitHub 仓库查看创作者相关仓库

下载

在 Manus 中运行

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.

同仓库更多 Skills

同仓库

paleo-auto

mocasus/paleo

Use when you want automatic token-saving without manual skill selection. Auto-detects long sessions (>15 turns), high token usage, or bulky context — then auto-enables paleo + paleo-trim-context + paleo-converse + paleo-budget at safe defaults. User does not need to think about which skill is active. Off: "disable paleo-auto" / "manual paleo".

2026-07-1217

paleo-budget

mocasus/paleo

Use when user says "budget", "token limit", "stay under N tokens", or wants per-task token caps. Enforce a hard token budget per task/response — track estimate, stop before limit, summarize if exceeded. Off: "no budget" / "unlimited".

2026-07-1017

paleo-converse

mocasus/paleo

Use when user says "compress conversation", "condense chat", "summarize history", "too long context", or a session has many old turns. Condense prior conversation turns into a tight summary, merge near-duplicate messages, keep the last N turns verbatim. Saves context tokens without losing task state. Off: "stop condensing" / "keep full history".

2026-07-1017

paleo-json

mocasus/paleo

Use when user says "compact json", "minify output", "tight structured", or the agent must emit JSON / structured data. Strip insignificant whitespace, collapse long arrays, shorten keys where safe, keep it parseable. Cuts tokens on structured payloads. Off: "pretty print" / "human-readable".

2026-07-1017

paleo-summary

mocasus/paleo

Use when user says "summarize output", "condense this", "tldr", or a tool result / log / file dump / long stdout is large. Reduce long content (tool output, logs, diffs, docs) to a concise intisari with key fields preserved. Saves context tokens on bulky payloads. Off: "full output" / "no summary".

2026-07-1017

paleo-trim-context

mocasus/paleo

Use when context window is large / token cost high / long session. Proactively trim, summarize, or drop stale content to save context tokens without losing the task state.

2026-07-1017

name	paleo
description	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".
version	2.4.1
license	MIT
metadata	{"hermes":{"tags":["tokens","compression","output","terse","efficiency"],"related_skills":["paleo-trim-context"]}}

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.