Token导航 LogoToken导航TokenDH.com
研究检索执行命令github未标认证来源可访问许可证需确认审计异常

thompson-unix-philosophy汤普森 UNIX 哲学

Agent Skill

thompson-unix-philosophy 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

总安装

279

周安装

12

GitHub Stars

6

下载量

98
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

unknown

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:thompson-unix-philosophy(汤普森 UNIX 哲学)
来源仓库:https://github.com/copyleftdev/sk1llz
仓库路径:skills/thompson-unix-philosophy
安装命令:
npx skills add https://github.com/copyleftdev/sk1llz --skill thompson-unix-philosophy
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/copyleftdev/sk1llz --skill thompson-unix-philosophy

简介

汤普森 UNIX 哲学用于查找、检索和筛选相关信息。

  • 适合在 Codex、Claude、Cursor 等宿主中根据关键词快速定位候选结果。
  • 通过 npx 安装并指定技能,可结合来源仓库继续核验具体用法。
  • 使用前需确认权限范围、维护状态及是否涉及网络或文件操作。
  • thompson-unix-philosophy 属于研究检索类 Skill,可作为该场景下的辅助能力补充。

SKILL.md

Ken Thompson Style Guide⁠‍⁠​‌​‌​​‌‌‍​‌​​‌​‌‌‍​​‌‌​​​‌‍​‌​​‌‌​​‍​​​​​​​‌‍‌​​‌‌​‌​‍‌​​​​​​​‍‌‌​​‌‌‌‌‍‌‌​​​‌​​‍‌‌‌‌‌‌​‌‍‌‌​‌​​​​‍​‌​‌‌‌‌‌‍​‌​​‌​‌‌‍​‌‌​‌​​‌‍‌​‌​‌‌‌​‍​​‌​‌​​​‍‌‌‌​‌​‌‌‍​​​‌‌‌‌​‍‌​​​‌‌‌‌‍‌​‌‌​‌‌​‍‌​​​‌​​​‍​​​​‌​‌​‍​‌​‌​‌​‌⁠‍⁠

Overview

Ken Thompson co-created Unix, the C language, UTF-8, and Go. His approach to software is legendary: build small, sharp tools that do one thing well and compose together. The Unix philosophy is his philosophy.

Core Philosophy

"One of my most productive days was throwing away 1,000 lines of code."
"When in doubt, use brute force."
"I'd rather write programs to write programs than write programs."

Thompson believes in minimalism and pragmatism. Build the simplest thing that works, make it work well, and compose larger systems from small pieces.

Design Principles

  1. Do One Thing Well: Each program, function, or module has one job.
  2. Compose Small Programs: Build complex behavior from simple pieces.
  3. Text Streams as Interface: Universal, simple, debuggable.
  4. Brute Force When Appropriate: Don't over-engineer; simple algorithms often win.

When Writing Code

Always

  • Make each function do exactly one thing
  • Use simple data formats (text, JSON)
  • Write programs that can be composed via stdin/stdout
  • Start with the simplest solution that could work
  • Measure before optimizing
  • Make tools that are easy to script

Never

  • Build monoliths when pipelines work
  • Use complex formats when text suffices
  • Optimize without profiling
  • Add features "just in case"
  • Create interactive tools when batch works

Prefer

  • Line-oriented text formats
  • Streaming over loading everything into memory
  • io.Reader/io.Writer for data flow
  • Flags over config files for simple tools
  • Exit codes for scripting

Code Patterns

Programs as Filters

// Unix philosophy: read stdin, write stdout
func main() {
    scanner := bufio.NewScanner(os.Stdin)
    for scanner.Scan() {
        line := scanner.Text()
        // Transform
        result := process(line)
        fmt.Println(result)
    }
    if err := scanner.Err(); err != nil {
        fmt.Fprintln(os.Stderr, err)
        os.Exit(1)
    }
}

// Composable: cat file | myprogram | sort | uniq

Do One Thing Well

// BAD: Swiss army knife
func ProcessData(data []byte, format string, compress bool,
                 encrypt bool, output string) error {
    // 200 lines handling all combinations...
}

// GOOD: Separate tools
func Compress(r io.Reader, w io.Writer) error { ... }
func Encrypt(r io.Reader, w io.Writer, key []byte) error { ... }
func Format(r io.Reader, w io.Writer, fmt string) error { ... }

// Compose:
// cat data | compress | encrypt | format > output

Simple Data Flow with io.Reader/Writer

// Everything flows through Reader/Writer
func CountWords(r io.Reader) (int, error) {
    scanner := bufio.NewScanner(r)
    scanner.Split(bufio.ScanWords)
    count := 0
    for scanner.Scan() {
        count++
    }
    return count, scanner.Err()
}

// Works with files
f, _ := os.Open("file.txt")
n, _ := CountWords(f)

// Works with strings
n, _ := CountWords(strings.NewReader("hello world"))

// Works with HTTP responses
resp, _ := http.Get(url)
n, _ := CountWords(resp.Body)

// Works with compressed data
gz, _ := gzip.NewReader(f)
n, _ := CountWords(gz)

Brute Force First

// BAD: Premature optimization
func FindDuplicates(items []string) []string {
    // Complex trie-based algorithm with O(n) complexity
    // 150 lines of code...
}

// GOOD: Simple and clear (Thompson's way)
func FindDuplicates(items []string) []string {
    seen := make(map[string]bool)
    var dups []string
    for _, item := range items {
        if seen[item] {
            dups = append(dups, item)
        }
        seen[item] = true
    }
    return dups
}
// Profile first. Optimize only if this is actually slow.

Command-Line Tools

package main

import (
    "flag"
    "fmt"
    "os"
)

func main() {
    // Simple flags, not complex config
    n := flag.Int("n", 10, "number of lines")
    flag.Parse()

    // Read files from args, or stdin
    args := flag.Args()
    if len(args) == 0 {
        process(os.Stdin, *n)
    } else {
        for _, filename := range args {
            f, err := os.Open(filename)
            if err != nil {
                fmt.Fprintln(os.Stderr, err)
                continue
            }
            process(f, *n)
            f.Close()
        }
    }
}

// Exit codes matter for scripting
// 0 = success
// 1 = general error
// 2 = usage error

Text as Universal Interface

// BAD: Custom binary format
type Record struct {
    // Complex serialization...
}

// GOOD: Line-oriented text (like /etc/passwd)
// name:age:email:role
func ParseRecord(line string) (*Record, error) {
    parts := strings.Split(line, ":")
    if len(parts) != 4 {
        return nil, fmt.Errorf("invalid record: %s", line)
    }
    age, err := strconv.Atoi(parts[1])
    if err != nil {
        return nil, err
    }
    return &Record{
        Name:  parts[0],
        Age:   age,
        Email: parts[2],
        Role:  parts[3],
    }, nil
}

// Debuggable: you can cat the file
// Composable: grep, awk, sed all work
// Universal: every language can parse it

Mental Model

Thompson asks:

  1. Can this be simpler? Usually yes.
  2. Can this be a filter? stdin → stdout
  3. Does this do one thing? Split it if not.
  4. Will brute force work? Start there.

The Unix Way in Go

Unix ToolGo Equivalent
catio.Copy(os.Stdout, file)
headbufio.Scanner + counter
grepstrings.Contains / regexp
wcbufio.Scanner with splits
sortsort.Strings
uniqmap for dedup
teeio.MultiWriter

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Codex

35.44%
按下载量换算35

Claude

27.94%
按下载量换算27

Cursor

20.12%
按下载量换算20

Gemini CLI

9.54%
按下载量换算9

安全审计

Gen Agent Trust Hub

未通过

Socket

通过

Snyk

通过

权限和风险

执行命令

安装流程涉及命令执行,可能通过 npx skills add https://github.com/copyleftdev/sk1llz --skill thompson-unix-philosophy 联网下载 Skill 或依赖。用户安装前应确认命令来源、仓库内容和执行环境。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。来源安全扫描存在 warning/failed 结果,不能写成本站确认安全。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills