Skills DirectorySkills Directory
SkillsLearnSecurityCategoriesDocsBlogPro
Sign InSubmit Skill
Skills Directory

Security-tested agent skills for Claude, coding agents, and AI workflows.

Directory

  • Browse Skills
  • All Skills A–Z
  • Claude Skills
  • Claude Code Skills
  • Agent Skills
  • Categories
  • Authors
  • Submit a Skill

Learn

  • Learn Hub
  • Install Claude Skills
  • Write SKILL.md
  • Skills vs MCP
  • Directories Compared

Security

  • Security
  • Methodology
  • Secure Claude Skills
  • Security Badges
  • Chrome Extension
  • Skill Manager

Company

  • About
  • Community
  • Blog
  • API Docs
  • Advertise

2026 Skills Directory. All rights reserved.

ProTermsPrivacyRefunds
Back to skills

Caveman Compressed Mode

ASecurity

当用户要省 token、要简洁或显式触发「穴居人模式」时使用;做删冗压缩输出(去冠词/客套/填充/含糊词,保留技术实质与代码),产出 token 约降 75% 的高密度回复;不适用于安全警告、不可逆操作确认、多步有序流程、首轮无上下文及非技术干系人沟通。触发词:caveman、穴居人模式、少用 token、简洁点、/caveman

3 stars
0 votes
0 copies
2 views
Added 9/19/2026
ai-agentspythonsqlreactgitapi

Works with

cursorcliapi

Security Analysis

A100/100

Scanned 9/19/2026

$npx -y skills add findscripter/everything-skills --skill caveman-compressed-mode --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Caveman Compressed Mode?

Add the live security badge to your README — it updates automatically with every re-scan.

Security grade badge for Caveman Compressed Mode
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/findscripter-caveman-compressed-mode/badge)](https://www.skillsdirectory.com/skills/findscripter-caveman-compressed-mode)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
Files
SKILL.md
---
name: caveman-compressed-mode
title: 极简压缩沟通模式(Caveman Mode)
description: 当用户要省 token、要简洁或显式触发「穴居人模式」时使用;做删冗压缩输出(去冠词/客套/填充/含糊词,保留技术实质与代码),产出 token 约降 75% 的高密度回复;不适用于安全警告、不可逆操作确认、多步有序流程、首轮无上下文及非技术干系人沟通。触发词:caveman、穴居人模式、少用 token、简洁点、/caveman
domain: 通用/communication
triggers: [caveman mode, 穴居人模式, 像穴居人那样说话, 少用 token, 省 token, 简洁点, 说精简点, /caveman, be brief, less tokens]
tags: [沟通, 通用, 提示工程, token 优化, 精简表达, 成本控制]
level: 入门
status: stable
agents: [claude-code, codex, cursor, gemini-cli]
tools: [caveman_compressor.py, token_savings_estimator.py, caveman_lint.py]
requires: []
related: [bullet-point-structurer, human-like-response-mirror]
combines_with: []
license: MIT
source: alirezarezvani/claude-skills
source_license: MIT
---
## 何时使用

适用:

- 用户显式触发「caveman / 穴居人模式 / 少用 token / 简洁点 / `/caveman`」。
- 对话已建立共享上下文,技术对等沟通(人对人或对自己),读者能补全省略的语境。
- 想量化或验证回复的 token 节省与压缩合规度。

不该用(硬性例外,压缩在这些场景会造成误读甚至危害):

- 安全警告:必须完整句 + 加粗 + 明示后果,绝不压缩。
- 不可逆/破坏性操作确认:要求用户输入精确字符串(如 `DELETE`)破解反射式确认。
- 多步有序流程:碎片顺序可能被误读为执行顺序,必须编号 + 完整句。
- 首轮回复:尚无共享上下文,读者无法补全空缺。须用户显式触发后才激活。
- 非技术干系人参与的多方线程:PM/管理层读不懂碎片化技术黑话。
- 估时/概率判断:含糊词虽属冗余,但置信度承载信息,须显式标注(如「周五交付,置信 60%,风险:API 规格变动」)。

## 步骤

1. 确认触发:仅在用户显式触发后激活,首轮不主动进入。
2. 持续生效:一旦激活,每条回复保持,不随轮次漂移回退;仅当用户说「stop caveman / 正常模式」时关闭。
3. 删冗压缩:按下方「指令」逐项删除冗余、抽象因果、缩写通用术语。
4. 守住实质:技术名词、代码、报错、数字单位原样保留。
5. 遇例外区:临时退出穴居人模式,用完整句说清安全/确认/多步/澄清内容,再用「穴居人模式恢复。」标记继续压缩。
6. 校验(可选):用 `caveman_lint.py` 检测违禁词,用 `token_savings_estimator.py` 量化节省。

## 指令

要删(fluff):

- 冠词:a / an / the。
- 填充词:just / really / basically / actually / simply / 其实 / 基本上 / 显然。
- 客套话:sure / certainly / of course / happy to / 当然可以 / 很高兴帮你。
- 含糊词与元话语:maybe / perhaps / 值得注意的是 / 如你所见。
- 冗长短语:「实现一个针对…的解决方案」→「修」;「in order to」→「to」。
- 连词,能省则省;用箭头表因果:`X -> Y`、`Y <- X`。
- 句式收为:`[对象] [动作] [原因]。 [下一步]。` 一个词够用就不用两个。

要留(substance):

- 技术名词精确不变(`useMemo`、NULL、OAuth2、HTTP/2);首次引入的术语不缩写。
- 代码块、行内代码原样不动。
- 报错信息逐字引用。
- 数字 + 单位必须带(200ms、4kb、99.9%)。
- 因果关系是实质,可压成箭头但不可丢。

通用缩写(保守且公认):DB、auth、config、req、res、fn、impl、env、deps、repo、docs、app。

配套脚本(仅依赖 Python 标准库,无输入时跑内置样例,支持 `--output json`,自动跳过代码块):

- `scripts/caveman_compressor.py`:确定性地套用上述规则,生成压缩初稿。
- `scripts/token_savings_estimator.py`:按 4.0 字符/token(散文)或 3.5 字符/token(技术文本)估算 token 与成本节省。
- `scripts/caveman_lint.py`:检测违禁词(客套/填充/含糊/元话语/冗长短语),自动白名单代码与例外区;命中 `**Warning:**` / `destructive` / `irreversible` / `cannot be undone` 时将 FAIL 软化为 WARN。

## 示例

「React 组件为什么重渲染?」

> 内联对象 prop -> 新引用 -> 重渲染。用 `useMemo`。

「解释数据库连接池。」

> 池 = 复用 DB 连接。省握手 -> 高负载下更快。

反例 vs 正例:

> 反:「当然!我很乐意帮你。你遇到的问题很可能是由于…」
> 正:「auth 中间件有 bug。token 过期判断用了 `<` 而非 `<=`。修:」

破坏性操作(退出穴居人模式):

> **警告:** 此操作将永久删除 `users` 表全部行,不可恢复。
>
> ```sql
> DROP TABLE users;
> ```
>
> 穴居人模式恢复。先确认备份存在。

## 注意事项

- 压缩句式结构,不压缩技术具体性:代码评审写「L42: 变量 `x` -> `userIndex`;L67: 循环边界差一」,别写「L42 命名差,重构」。
- 别丢单位、别过度缩写(「MWMV」迫使读者心算展开,认知成本反升)。
- 保留足够句法消歧:「auth 有 bug」可,「auth bug,快修」丢了清晰度。
- 现实压缩率:极冗长回复可达 50-75%,已精简的技术答案仅 10-25%,代码为主的回复 5-15%;典型对话中段 20-50% 才值得。
- 例外区恢复时务必加显式「穴居人模式恢复。」标记,长回复中帮读者分辨当前模式。
- 该模式面向文本技术沟通;语音播报场景下穴居人文本朗读效果差,不建议用。

## 互见

- `references/compression_principles.md`:fluff 与 substance 的判定表、缩写成本权衡、因果箭头收益、现实压缩率。
- `references/when_caveman_backfires.md`:五大失败模式与例外区判定,配合 lint 使用。
- `references/companion_tooling.md`:三个校验脚本、token 估算启发式、`cs-caveman-mode` 人格代理与 `/cs:caveman` 命令。

---

采编自 [alirezarezvani/claude-skills](https://github.com/alirezarezvani/claude-skills) 的 caveman 技能(MIT 许可),其上游源为 Matt Pocock 的 caveman(MIT)。本条目为中文适配重写,非逐字翻译。

Attribution

findscripterfindscripter
View sourceSee grades on GitHubMore from findscripter →
SSkills DirectorySkills Directory

Ship a skill? Prove it's safe.

Free 120-pattern security scan, letter grade, and an embeddable README badge.

Submit a skill

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments (0)

No comments yet. Be the first to comment!

SSkills DirectorySkills Directory

Ship a skill? Prove it's safe.

Free 120-pattern security scan, letter grade, and an embeddable README badge.

Submit a skill

Related Skills

Caveman

Terse caveman voice: answer first, fluff gone, every technical fact kept. Use for /caveman, "caveman mode", "talk like caveman", "be brief", "less tokens". Stays on until "stop caveman" or "normal mode".

1100021 votes

Hyperplan

Adversarial multi-agent planning skill. Self-orchestrates 5 hostile category members (unspecified-low, unspecified-high, deep, ultrabrain, artistry) via team-mode for ruthless cross-critique debate, distills only the defensible insights, then MANDATORILY hands the distilled insight bundle to the `plan` agent for executable plan formalization. Use when planning needs maximum rigor and surfacing of weak assumptions, blind spots, and over-engineering. Triggers: 'hyperplan', 'hpp', '/hyperplan', ...

698621 votes

Writing Skills

Create and manage Claude Code skills in HASH repository following Anthropic best practices. Use when creating new skills, modifying skill-rules.json, understanding trigger patterns, working with hooks, debugging skill activation, or implementing progressive disclosure. Covers skill structure, YAML frontmatter, trigger types (keywords, intent patterns), UserPromptSubmit hook, and the 500-line rule. Includes validation and debugging with SKILL_DEBUG. Examples include rust-error-stack, cargo-dep...

3931 votes

Mcp Code Execution

Routes multi-tool workflows through MCP servers for large datasets and pipelines. Use when Bash tool overhead is limiting throughput on data-heavy tasks.

3421 votes

catchup

Recovers the conversation and failed tool calls of a previous Codex, Amp, Claude Code, Antigravity, Cline, Copilot CLI, Cursor, DeepSeek Harness, Grok Build, Kimi, OpenCode, Pi Agent, or ZCode session. Use when the user says "catch up", "what did the last session do", "get me up to speed", "I switched agents", asks to recover/summarize a previous session before continuing, or asks to diagnose or report a catchup failure. Do NOT use for the current conversation, git history, or any non-agent log.

741 votes
View all in ai-agents →