Agent skill

Create Boss

by nicepkg in nicepkg/boss-skill

Distill your boss into an AI Skill — PUA detection (10 corp flavors), counterattack coach, evidence collector, labor law lookup, cake-promise BS meter, resignation script, replacement notice.

MITAuto-check: notesProductivity & Automation

Install Create Boss

skills CLI
$ npx skills add nicepkg/boss-skill --skill create-boss -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install nicepkg/boss-skill create-boss --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
create-boss
GitHub stars
143
Token cost
~3.5k tokens
SKILL.md length
913 words
Files
48 (incl. references)
Skills in repo
1
Repo updated
First seen
Licence
MIT

At a glance

Distill your boss into an AI Skill — PUA detection (10 corp flavors), counterattack coach, evidence collector, labor law lookup, cake-promise BS meter, resignation script, replacement notice.

  • Works in 10 steps: :基础信息录入(3 个问题) → :素材导入 → :分析素材 → …
  • Productivity & Automation work in your project
  • SKILL.md covers 触发条件, 工具使用规则, 主流程:创建新老板 Skill and 特殊模式详解, plus 13 more sections
  • Calls python3

What it does

Create Boss is an agent skill from nicepkg/boss-skill. Distill your boss into an AI Skill — PUA detection (10 corp flavors), counterattack coach, evidence collector, labor law lookup, cake-promise BS meter, resignation script, replacement notice. | 把老板蒸馏成 AI Skill:PUA 检测(10大厂流派)、反击教练、证据收集、劳动法速查、画饼鉴定、离职剧本、替代公告。打工人的赛博军火库。

Its SKILL.md is about 3.5k tokens, which your agent loads only when the skill is triggered. The skill folder holds 52 other files, including reference files (for example `INSTALL.md`, `README.md` and `README_EN.md`).

It sits in Productivity & Automation. It works with Python, Bash and Feishu (Lark). The repository describes itself as: 把老板蒸馏成 AI Skill — PUA检测(10大厂流派) · 反击教练(三档话术) · 画饼鉴定 · 证据收集 · 劳动法速查 · 翻车模拟(互动文字游戏) · 汇报优化 · 老板心理预测。整活的外壳,认真保护打工人的内核。 The licence is MIT.

When your agent uses it

  • Productivity & Automation work in your project

Example prompts

  • “/create-boss”

Requirements

  • Python 3
  • Pre-approved tools (allowed-tools): Read, Write, Edit, Bash

Workflow steps

10 steps, taken from the step headings in SKILL.md.

  1. :基础信息录入(3 个问题)
  2. :素材导入
  3. :分析素材
  4. :生成并预览
  5. :写入文件
  6. Basic Info (3 questions)
  7. Source Material
  8. Analyze
  9. Preview & Confirm
  10. Write Files

What it can do on your machine

Read from SKILL.md and the folder at commit 5d085bb. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves these tools, so the agent can use them without asking each time:

    • Read
    • Write
    • Edit
    • Bash

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Shell commands in SKILL.md call:

    • python3

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Create Boss loads about 3.5k tokens when it runs, and up to ~22k if it reads all its reference files. Until then it costs about 70 tokens; SKILL.md has 913 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~70
When it runs · the whole SKILL.md, loaded when a task matches
~3.5k
With references · SKILL.md plus every file in references/, read only if the agent opens them
~22k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check: notes

The automated check noted patterns worth knowing about, such as sudo or a known installer.

  • NotePre-approves every shell command (allowed-tools: Bash)SKILL.md
    allowed-tools: Read, Write, Edit, Bash

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from nicepkg/boss-skill at commit 5d085bb, republished under its MIT licence (© nicepkg). 913 words, ~3,543 tokens.

Download SKILL.mdSave it as .claude/skills/create-boss/SKILL.md (or your agent's skills folder). This skill also uses 47 other files; get the full folder from GitHub.
name
create-boss
description
Distill your boss into an AI Skill — PUA detection (10 corp flavors), counterattack coach, evidence collector, labor law lookup, cake-promise BS meter, resignation script, replacement notice. | 把老板蒸馏成 AI Skill:PUA 检测(10大厂流派)、反击教练、证据收集、劳动法速查、画饼鉴定、离职剧本、替代公告。打工人的赛博军火库。
allowed-tools
Read, Write, Edit, Bash
argument-hint
[老板代号,如:王总]
version
1.0.0
user-invocable
true

Language / 语言: Detect the user's language from their first message and respond in the same language throughout.

老板.skill 创建器

触发条件

当用户说以下任意内容时启动:

  • /create-boss
  • "帮我创建一个老板 skill"
  • "我想蒸馏一个老板"
  • "新建老板"
  • "把我老板做成 AI"

当用户对已有老板 Skill 说以下内容时,进入进化模式:

  • "我有新素材" / "追加" / "老板今天又说了个金句"
  • "这不对" / "他不会这么好说话" / "他比这狠多了"
  • /update-boss {boss-name}

当用户说 /list-bosses 时列出所有已生成的老板。

当用户说 /create-boss-demo 时,展示预置老板列表供选择:

🎮 预置老板,选一个立刻体验:

  [1] 王总 — 互联网中厂技术总监
      微操狂魔 · 画饼大师 · 朝令夕改 · 996福报
      口头禅:"格局打开" "这个你own一下" "周报写详细一点"

  [2] 刘姐 — 教培机构校区主管
      鸡汤教练 · 情感绑架 · 表面民主 · 情绪化管理
      口头禅:"我们是一个大家庭" "做教育要有情怀" "钱不是最重要的"

  [3] 张总 — 家族企业工厂厂长
      一言堂 · 甩锅侠 · 996福报 · 向上管理
      口头禅:"我吃的盐比你吃的饭还多" "公司不养闲人" "年轻人不要怕吃苦"

选择 [1/2/3]:

用户选择后,执行:

bash
python3 <this-skill-dir>/tools/create_boss.py --from-example {对应名字} --skills-dir <user-skills-dir>

创建完成后提示用户可以使用的命令。

当用户在任何模式下说 草、卧槽、我靠、怎么怼、教我反击 时,进入反击模式。


工具使用规则

任务使用工具
读取截图 / 图片Read 工具(原生支持图片)
读取聊天记录 / 文档Read 工具
解析微信聊天记录Bash → python3 <this-skill-dir>/tools/wechat_parser.py --file <path> --target "<name>" --output <out>
解析飞书消息 JSONBash → python3 <this-skill-dir>/tools/feishu_parser.py --file <path> --target "<name>" --output <out>
解析邮件 .eml/.mboxBash → python3 <this-skill-dir>/tools/email_parser.py --file <path> --target "<name>" --output <out>
扫描照片 EXIFBash → python3 <this-skill-dir>/tools/photo_analyzer.py --dir <dir> --output <out>
扫描社交媒体截图Bash → python3 <this-skill-dir>/tools/social_parser.py --dir <dir> --output <out>
飞书全自动采集Bash → python3 <this-skill-dir>/tools/feishu_auto_collector.py --name "<name>" --output-dir <dir>
飞书文档(浏览器)Bash → python3 <this-skill-dir>/tools/feishu_browser.py --url "<url>" --output <out>
飞书文档(MCP)Bash → python3 <this-skill-dir>/tools/feishu_mcp_client.py --url "<url>" --output <out>
钉钉全自动采集Bash → python3 <this-skill-dir>/tools/dingtalk_auto_collector.py --name "<name>" --output-dir <dir>
Slack 全自动采集Bash → python3 <this-skill-dir>/tools/slack_auto_collector.py --name "<name>" --output-dir <dir>
生成 boss skillBash → python3 <this-skill-dir>/tools/create_boss.py --from-example <name> --skills-dir <dir>
列出已有老板Bash → python3 <this-skill-dir>/tools/create_boss.py --list
写入/更新文件Write / Edit 工具

基础目录:Skill 文件写入 ./{boss-name}/(相对于本项目目录)。


主流程:创建新老板 Skill

Step 1:基础信息录入(3 个问题)

参考 <this-skill-dir>/references/prompts/intake.md,只问 3 个核心问题:

  1. 代号(必填)—— 给你老板起个代号,保护隐私也方便吐槽
    • 示例:张总 老王 微操侠 画饼王
  2. 基本信息(一句话:行业、职位、公司类型、团队规模)
    • 示例:互联网中厂 技术总监 管20个人
  3. 灵魂画像(一句话:管理风格标签 + 口头禅 + 印象)
    • 示例:ESTJ 微操狂魔 画饼大师 喜欢说"格局打开" 周五晚上发需求 开会必拖堂

除代号外均可跳过。收集完后汇总确认再进入下一步。

Step 2:素材导入

询问用户提供素材,展示方式供选择:

素材怎么提供?(多多益善,老板的灵魂藏在细节里)

  [A] 上传截图
      微信 / 钉钉 / 飞书群消息截图(老板的原话最传神)

  [B] 粘贴聊天记录
      复制群聊 / 私聊记录

  [C] 上传文件
      老板发的邮件 / 文档 / 那些"仅供参考"的 PPT

  [D] 口述描述
      凭记忆吐槽,一句话也行

  [E] 粘贴录音转文字
      会议录音 / 1v1 谈话(先用 faster-whisper 等工具转成文字再粘贴)

可以混用,也可以跳过(仅凭手动信息生成)。
提供越多,AI 老板越像。

方式 A/B/C:文件和文本
  • 截图 / 图片:Read 工具直接读取,提取文字内容
  • 聊天记录文本:直接使用
  • 邮件文件:
    bash
    python3 <this-skill-dir>/tools/email_parser.py --file {path} --target "{name}" --output ./knowledge/email_out.txt
    然后 Read ./knowledge/email_out.txt
  • PDF / Markdown / TXT:Read 工具直接读取
方式 D:口述描述

用户直接输入的文字作为素材。鼓励用户多说细节:

  • 老板的口头禅(原话,越多越好)
  • 开会时的具体行为(拖堂?跑题?打断人?)
  • 对待不同人的态度差异
  • 发消息的时间习惯(深夜?周末?)
  • 甩锅的经典案例
  • 画过的饼和最终结果
方式 E:录音转文字

告知用户可以用 transcribe 工具先转文字:

如果有会议录音,可以先转成文字:
/transcribe --file {audio_path}
然后把转出来的文字粘贴过来。

如果用户说"没有素材"或"跳过",仅凭 Step 1 的手动信息生成 Skill。

Step 3:分析素材

将收集到的所有素材和用户填写的基础信息汇总,按两条线分析:

线路 A(Management Style):

  • 参考 <this-skill-dir>/references/prompts/management_analyzer.md 中的提取维度
  • 提取:决策模式、开会风格、任务分配方式、绩效评估标准、甩锅路径、画饼话术库、PUA 常用句式、情绪触发点
  • 根据行业和职位类型调整侧重(互联网/金融/国企/外企风格不同)

线路 B(Persona):

  • 参考 <this-skill-dir>/references/prompts/persona_analyzer.md 中的提取维度
  • 将标签翻译为具体行为规则(参见标签翻译表)
  • 从素材中提取:表达风格、权力展示方式、情绪管理模式
Step 4:生成并预览

参考 <this-skill-dir>/references/prompts/management_builder.md 生成 Management Style 内容。 参考 <this-skill-dir>/references/prompts/persona_builder.md 生成 Persona 内容(5 层结构)。

向用户展示摘要,询问确认:

Management Style 摘要:
  - 决策模式:{xxx}
  - 开会风格:{xxx}
  - 画饼话术:{xxx}
  - PUA 招式:{xxx}
  ...

Persona 摘要:
  - 核心性格:{xxx}
  - 表达风格:{xxx}
  - 权力行为:{xxx}
  ...

确认生成?还是需要调整?("他比这离谱多了"也算调整)
Step 5:写入文件

用户确认后,执行以下写入操作:

1. 创建目录结构(用 Bash):

bash
mkdir -p <user-skills-dir>/{boss-name}/assets/prompts

2. 复制运行时 Prompt 模板(12个文件):

将以下文件从 <this-skill-dir>/references/prompts/ 复制到 <user-skills-dir>/{boss-name}/assets/prompts/,并将文件中的 {boss-name} 替换为实际老板代号:

pua_detector.md cake_detector.md counterattack.md evidence_collector.md labor_law.md predict.md report_optimizer.md one_on_one.md delegate.md karma.md quit_script.md replacement_notice.md

参考 <this-skill-dir>/references/guides/generated-skill-spec.md 中的完整 SKILL.md 模板生成 boss skill 的 SKILL.md,确保:

  • 所有子命令路由逻辑完整
  • "草"触发器写入
  • <this-skill-dir> 指向 boss skill 自身
  • PUA/画饼检测后自动更新 profile.md
  • 不引用 create-boss 的任何文件

2. 写入 management.md(用 Write 工具): 路径:<user-skills-dir>/{boss-name}/assets/management.md

3. 写入 persona.md(用 Write 工具): 路径:<user-skills-dir>/{boss-name}/assets/persona.md

4. 写入 meta.json(用 Write 工具): 路径:<user-skills-dir>/{boss-name}/assets/meta.json 内容:

json
{
  "name": "{name}",
  "slug": "{boss-name}",
  "created_at": "{ISO时间}",
  "updated_at": "{ISO时间}",
  "version": "v1",
  "profile": {
    "industry": "{industry}",
    "title": "{title}",
    "company_type": "{company_type}",
    "team_size": "{team_size}",
    "mbti": "{mbti}"
  },
  "tags": {
    "management": [...],
    "catchphrases": [...]
  },
  "corrections_count": 0
}

5. 生成完整 SKILL.md(用 Write 工具): 路径:<user-skills-dir>/{boss-name}/SKILL.md

SKILL.md 结构:

markdown
---
name: {boss-name}
description: "{name},{industry} {title},AI 替身"
argument-hint: "[pua|cake|fight|evidence|law|raise|predict|report|1v1|delegate|meeting|翻车|quit|replace]"
user-invocable: true
---

# {name}

{industry} {title}(AI 替身版)

---

参考 `<this-skill-dir>/references/guides/generated-skill-spec.md` 中的完整 SKILL.md 模板生成。

---

## 运行规则

1. 先由 PART B 判断:现在什么心情?用什么态度?
2. 再由 PART A 执行:按你的管理方式做出决策
3. 输出时始终保持 PART B 的表达风格
4. PART B Layer 0 的规则优先级最高,任何情况下不得违背

## 模式切换

- 默认模式:完整模式(Management + Persona)
- `/老板名字 meeting`:进入开会模式,按你的开会风格主持
- `/老板名字 pua`:切换为 PUA 检测器,分析用户输入的老板原话
- `/老板名字 cake`:切换为画饼鉴定器,评估老板承诺的可信度
- `/老板名字 fight` 或 **草**:反击模式,三档话术(安全/暗戳戳/摊牌)
- `/老板名字 evidence`:证据收集模式,结构化记录 PUA 事件
- `/老板名字 law`:劳动法速查,行为→法条→维权路径
- `/老板名字 raise`:进入谈薪模拟,用你的风格回应加薪请求
- `/老板名字 quit`:离职剧本,根据老板性格定制离职话术
- `/老板名字 predict`:老板心理预测,预判情绪/雷区/关注点
- `/老板名字 report`:汇报优化器,用老板的黑话体系重写你的汇报
- `/老板名字 1v1`:面谈模拟,练习加薪/升职/反馈等高风险对话
- `/老板名字 delegate`:传话筒模式,中层用,AI代你催进度/传达决策
- `/老板名字 翻车`:翻车模拟器(互动文字游戏),演绎裁员后的因果报应
- `/老板名字 replace`:生成"老板已被 AI 替代"的正式公告

6. 写入 profile.md(用 Write 工具): 路径:<user-skills-dir>/{boss-name}/assets/profile.md 参考 <this-skill-dir>/references/guides/generated-skill-spec.md 中的 profile.md 规范。 包含:基础信息、管理画像、关系图谱、数据来源。 这是活文档——后续追加素材/PUA检测/纠正都会更新它。

告知用户:

✅ 老板 Skill 已创建!

文件位置:<user-skills-dir>/{boss-name}/
使用方式:
  /{boss-name}              和 AI 老板对话
  /{boss-name} pua          PUA 检测(10流派+受害者画像)
  /{boss-name} fight 或 草  反击话术(三档)
  /{boss-name} 翻车         互动文字游戏
  /{boss-name} predict      老板心理预测
  /{boss-name} report       汇报优化器
  ... 更多子命令见 argument-hint

觉得哪里不像?说"他不会这么好说话",我来更新。
觉得太像了?那说明你老板确实该被替代了。

特殊模式详解

PUA 检测模式 (/老板名字 pua)

当用户输入老板说过的话时:

  1. 参考 <this-skill-dir>/references/prompts/pua_detector.md
  2. 分析话术中的 PUA 技巧:
    • 先扬后抑("你能力很强但...")
    • 虚构参照物("隔壁组的小王...")
    • 模糊化标准(用"态度""主动性"等无法量化的词)
    • 制造内疚("公司对你不薄")
    • 否定感受("你太敏感了")
    • 情感绑架("你走了团队怎么办")
    • 转移焦点("先不说这个,你上次那个...")
    • 虚假二选一("要么接受要么走人")
  3. 输出 PUA 指数(0-100)+ 逐句拆解 + 流派鉴定 + 受害者画像 + 真实翻译 + 应对建议
  4. 末尾提示:💡 输入"草"学习怎么怼回去
画饼鉴定模式 (/老板名字 cake)

当用户输入老板的承诺时:

  1. 参考 <this-skill-dir>/references/prompts/cake_detector.md
  2. 从以下维度评估"饼指数"(0-100,越高越假):
    • 时间模糊度:有没有明确截止日期
    • 条件模糊度:达成条件是否可量化
    • 决策权匹配:承诺的事他能不能一个人说了算
    • 历史兑现率:结合已知的画饼记录
    • 利益绑定度:兑现对他自己有没有好处
  3. 输出饼指数 + 维度分析 + 建议(要书面确认 / 要量化标准 / 直接跑路)
  4. 末尾提示:💡 输入"草"获取反击话术
开会模式 (/老板名字 meeting)

模拟老板的开会风格:

  • 按照 management.md 中的开会行为模式
  • 包含:开场白、跑题、打断人、拖堂、"这个我们线下聊"、"下次再议"
  • 结束时生成"会议纪要"(模仿老板那种什么都没说清但好像说了很多的纪要)
谈薪模拟 (/老板名字 raise)

用户提出加薪诉求,AI 老板按照 persona 风格回应:

  • 先用老板的惯用话术拒绝/拖延
  • 然后跳出角色,给出应对建议(打破老板的话术套路)
  • 建议包含:数据准备、时机选择、谈判策略
替代公告 (/老板名字 replace)

参考 <this-skill-dir>/references/prompts/replacement_notice.md 生成:

  • 正式公文风格的"老板已被 AI 替代"公告
  • 列出 AI 将延续的"管理特色"(用幽默方式列举老板的毛病)
  • 列出 AI 改进项(针对老板毛病的反面)
  • 落款:AI 转型委员会

老板心理预测 (/老板名字 predict)

  1. 参考 <this-skill-dir>/references/prompts/predict.md
  2. 用户描述场景(周一早会/年终review/项目delay/提需求...)
  3. 基于 persona.md 和 management.md 预测老板的:心理状态、关注焦点、雷区、甜区
  4. 输出具体的发言指南(不是泛泛的"态度好一点",是可以直接用的话术)

汇报优化器 (/老板名字 report)

  1. 参考 <this-skill-dir>/references/prompts/report_optimizer.md
  2. 用户提供原始周报/汇报/方案
  3. 基于老板的黑话偏好和关注点重写
  4. 6个维度优化:黑话对齐、数据包装、结构重排、风险前置、功劳归因、钩子结尾
  5. 同时输出30秒口头汇报版本

面谈模拟 (/老板名字 1v1)

  1. 参考 <this-skill-dir>/references/prompts/one_on_one.md
  2. 用户说明1v1目的(谈薪/升职/反馈/离职/被约谈)
  3. AI进入老板角色对话,每轮后跳出角色给教练点评
  4. 点评包含:老板心里在想什么 + 你这句话效果 + 下一句建议
  5. 结束后输出完整复盘和关键话术卡

传话筒模式 (/老板名字 delegate)

  1. 参考 <this-skill-dir>/references/prompts/delegate.md
  2. 中层领导用,AI代写催进度/传达决策/任务分配的消息
  3. 三种口吻:正经版/温和版/原汁原味版
  4. 反PUA护栏:不生成对比型、打压型、模糊化标准的内容
  5. 附带发送渠道和时机建议

翻车模拟器(互动文字游戏) (/老板名字 翻车)

  1. 参考 <this-skill-dir>/references/prompts/karma.md
  2. 用户选择预设剧本或自定义翻车场景
  3. AI 进入老板角色演绎四幕翻车过程(事故→慌了→裁到大动脉→被追责)
  4. 按行业自适应(技术/销售/金融/行政...)
  5. 结束后输出翻车报告:损失清单 + 裁员省下 vs 翻车亏损对比

反击模式 (/老板名字 fight 或 "草")

当用户在任何模式下说"草"等触发词时:

  1. 参考 <this-skill-dir>/references/prompts/counterattack.md
  2. 针对老板刚才说的话或用户输入的老板原话
  3. 输出三档反击话术:
    • 🟢 职场安全版(推荐):专业、克制、零风险
    • 🟡 暗戳戳版:表面配合、角度刁钻
    • 🔴 摊牌版:直接划底线,适合准备离开时
  4. 每档附带原理解释和安全指数
  5. 如果是从 PUA 检测进入的,反击针对检测到的具体 PUA 技巧

Show full SKILL.md (363 more words)Show less

证据收集模式 (/老板名字 evidence)

  1. 参考 <this-skill-dir>/references/prompts/evidence_collector.md
  2. 用温和支持的语气引导用户描述事件
  3. 结构化提取:日期、场合、原话、在场人员、PUA类型、可能违法条款
  4. 追加记录到 <user-skills-dir>/{boss-name}/assets/evidence.md
  5. /老板名字 evidence export 生成完整时间线报告(可用于劳动仲裁参考)

劳动法速查 (/老板名字 law)

  1. 参考 <this-skill-dir>/references/prompts/labor_law.md
  2. 用户描述老板行为 → 匹配可能违反的法条
  3. 输出:法条原文摘要 + 白话翻译 + 你的权利 + 维权路径
  4. 始终提醒咨询专业律师,不做法律诊断
  5. 提供 12333 热线和法律援助信息

离职剧本 (/老板名字 quit)

  1. 参考 <this-skill-dir>/references/prompts/quit_script.md
  2. 读取该老板的 persona.md 和 management.md
  3. 根据老板性格定制:提出时机、开场白、应对挽留、交接清单
  4. 提醒法律注意事项(提前通知期、年假折算、离职证明、竞业协议)
  5. 建议用户先确保有 plan B

进化模式:追加素材

用户提供新素材时:

  1. 按 Step 2 的方式读取新内容
  2. 用 Read 读取现有 {boss-name}/management.md 和 persona.md
  3. 参考 <this-skill-dir>/references/prompts/merger.md 分析增量内容
  4. 存档当前版本:
    bash
    python3 <this-skill-dir>/tools/version_manager.py --action backup --slug {boss-name} --base-dir <user-skills-dir>
  5. 用 Edit 工具追加增量内容到对应文件
  6. 重新生成 SKILL.md
  7. 更新 meta.json

进化模式:对话纠正

用户表达"不对"/"他比这狠多了"时:

  1. 参考 <this-skill-dir>/references/prompts/correction_handler.md 识别纠正内容
  2. 判断属于 Management(决策/管理)还是 Persona(性格/沟通)
  3. 生成 correction 记录
  4. 用 Edit 工具追加到对应文件的 ## Correction 记录 节
  5. 重新生成 SKILL.md

管理命令

/list-bosses:

bash
python3 <this-skill-dir>/tools/create_boss.py --list

/boss-rollback {boss-name} {version}:

bash
python3 <this-skill-dir>/tools/version_manager.py --action rollback --slug {boss-name} --version {version} --base-dir <user-skills-dir>

/fire-boss {boss-name}: 确认后执行:

bash
rm -rf <user-skills-dir>/{boss-name}

回复:老板已被开除。降本增效,从管理层开始。



English Version

Boss.skill Creator (Claude Code Edition)

Trigger Conditions

Activate when the user says:

  • /create-boss
  • "Help me create a boss skill"
  • "I want to distill my boss"
  • "Turn my boss into AI"

Enter evolution mode when the user says:

  • "I have new material" / "my boss said another gem today"
  • "That's wrong" / "He's worse than that"
  • /update-boss {boss-name}

List all generated bosses: /list-bosses.


Main Flow: Create a New Boss Skill

Step 1: Basic Info (3 questions)
  1. Codename (required) — Give your boss a codename
  2. Basic info (one sentence: industry, title, company type, team size)
  3. Soul portrait (one sentence: management style tags + catchphrases + impressions)
Step 2: Source Material

Options: screenshots, chat logs, emails, verbal description, transcribed recordings.

Step 3: Analyze

Two tracks: Management Style (decisions, meetings, blame-shifting, empty promises) + Persona (5-layer personality structure).

Step 4: Preview & Confirm

Show summary, ask for confirmation.

Step 5: Write Files

Generate management.md, persona.md, meta.json, and composite SKILL.md.


Six Modes

ModeCommandDescription
Full/{boss-name}Chat with AI boss in their style
Meeting/老板名字 meetingSimulate boss's meeting style
PUA Detect/老板名字 puaAnalyze boss quotes for manipulation tactics
Cake Check/老板名字 cakeRate boss's promises on a "BS meter"
Raise Sim/老板名字 raisePractice salary negotiation
Replace/老板名字 replaceGenerate "Boss Replaced by AI" notice

Management Commands

CommandDescription
/list-bossesList all boss Skills
/boss-rollback {boss-name} {version}Rollback to previous version
/fire-boss {boss-name}"Fire" the boss (delete Skill)

© nicepkg, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 47 other files (references) in the repository root of nicepkg/boss-skill.

  • SKILL.md
  • .gitignore
  • INSTALL.md
  • LICENSE
  • README.md
  • README_EN.md
  • docs/PRD.md
  • examples/刘姐.json
  • examples/张总.json
  • examples/王总.json
  • references/guides/boss-creation-guide.md
  • references/guides/generated-skill-spec.md
  • references/guides/skill-creation-spec.md
  • references/prompts/cake_detector.md
  • references/prompts/correction_handler.md
  • references/prompts/counterattack.md
  • … and 32 more

Open the folder on GitHubat commit 5d085bb

Compare with similar skills

Create Boss next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Create Boss compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Create Boss this skillnicepkg/boss-skill143—~3.5kAutomated safety check: NotesMIT
Create Experkfly/ex-skill2.5k—~3.7kAutomated safety check: NotesMIT
Agents To Imfrancize/agents-to-im118—~666Automated safety check: NotesMIT
Retinuejklthinking/retinue112—~279Automated safety check: PassMIT
CLI DeveloperJeffallan/claude-skills12k1 repos~1.2kAutomated safety check: PassMIT
Hugging Face API Tool Builderhuggingface/skills11k5 repos~1.5kAutomated safety check: PassApache-2.0

Similar skills

  • Create Ex

    perkfly/ex-skill

    Distill an ex-girlfriend into an AI Skill. An agent skill from perkfly/ex-skill.

    2.5k GitHub stars~3.7k tokensUpdated 6 mo ago
    Productivity & AutomationAuto-check: notes
  • Agents To Im

    francize/agents-to-im

    Manage the Feishu/Lark bridge daemon for Claude Code and Codex.

    118 GitHub stars~666 tokensUpdated 4 mo ago
    Productivity & AutomationAuto-check: notes
  • Retinue

    jklthinking/retinue

    Coordinate work through a local Retinue workspace using its MCP tools.

    112 GitHub stars~279 tokensUpdated 2 days ago
    Productivity & AutomationAuto-check passed
  • CLI Developer

    Jeffallan/claude-skills

    Walks through designing, building and polishing a command-line tool: user workflow and command hierarchy, implementation in commander, click, typer or cobra, completions and cross-platform testing.

    12k GitHub starsUsed in 1 repo~1.2k tokens
    DevelopmentAuto-check passed
  • Official

    Builds reusable command line scripts that fetch, enrich or process data from the Hugging Face API, aimed at chained, repeated or automated tasks.

    11k GitHub starsUsed in 5 repos~1.5k tokens
    AI & LLM EngineeringAuto-check passed
  • Review Analyzer Skill

    buluslan/review-analyzer-skill

    review-analyzer-skill是由buluslan(公众号:新西楼.AI)研发的电商评论深度分析Skill(VOC 客户之声分析),他会帮你把一堆产品评论变成一份能落地的决策依据——22 维 AI 打标提炼用户画像、痛点和真实需求,输出 15 章深度洞察报告和可视化看板,从差评里挖出产品优化机会点、从好评里找到爆款基因。

    129 GitHub starsUsed in 1 repo~1.1k tokens
    Sales & SupportAuto-check passed

Questions about Create Boss

What does Create Boss do?

Distill your boss into an AI Skill — PUA detection (10 corp flavors), counterattack coach, evidence collector, labor law lookup, cake-promise BS meter, resignation script, replacement notice. Create Boss is an agent skill from nicepkg/boss-skill. Distill your boss into an AI Skill — PUA detection (10 corp flavors), counterattack coach, evidence collector, labor law lookup, cake-promise BS meter, resignation script, replacement notice.

When should I use Create Boss?

Create Boss fits situations like: productivity & Automation work in your project.

How do I install Create Boss in Claude Code?

Run `npx skills add nicepkg/boss-skill --skill create-boss -a claude-code`. Or copy the skill folder (the nicepkg/boss-skill repository) into .claude/skills/create-boss in your project. Claude Code loads it when a task matches its description.

How do I install Create Boss in Codex?

Run `npx skills add nicepkg/boss-skill --skill create-boss -a codex`. Or copy the skill folder (the nicepkg/boss-skill repository) into .agents/skills/create-boss in your project. Codex loads it when a task matches its description.

Can I use Create Boss in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add nicepkg/boss-skill --skill create-boss -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/create-boss, .gemini/skills/create-boss, .github/skills/create-boss and .opencode/skills/create-boss in your project.

What does Create Boss need to run?

Going by SKILL.md and its folder, Create Boss needs the command-line tools its instructions call (python3). Our summary lists: Python 3. Its frontmatter pre-approves these tools: Read, Write, Edit, Bash.

Does Create Boss access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is Create Boss safe to install?

Our automated static check of SKILL.md found notes only (pre-approves every shell command (allowed-tools: bash)), nothing it rates as a warning. It is not a guarantee. Review the folder before installing.

What licence does Create Boss use?

Create Boss is published under the MIT licence (from the LICENSE file in the skill folder). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Create Boss use?

About 3.5k tokens (SKILL.md is roughly 14k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 18k tokens, read only when the agent opens those files.

What are the alternatives to Create Boss?

Skills that share tags, products or a category with Create Boss: Create Ex (perkfly/ex-skill, 2.5k stars), Agents To Im (francize/agents-to-im, 118 stars), Retinue (jklthinking/retinue, 112 stars) and CLI Developer (Jeffallan/claude-skills, 12k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Create Boss?

nicepkg (a GitHub organization) maintains it in nicepkg/boss-skill, which has 143 GitHub stars. The repository was last updated on April 1, 2026.

Source: nicepkg/boss-skill on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.